This guide is intended for administrators who need to set up, configure, and maintain clusters with SUSE® Linux Enterprise High Availability Extension. For quick and efficient configuration and administration, the product includes both a graphical user interface and a command line interface (CLI). For performing key tasks, both approaches are covered in this guide. Thus, you can choose the appropriate tool that matches your needs.
conntrackd
amsterdam
)drbdmon
drbdmon
Copyright © 2006–2019 SUSE LLC and contributors. All rights reserved.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or (at your option) version 1.3; with the Invariant Section being this copyright notice and license. A copy of the license version 1.2 is included in the section entitled “GNU Free Documentation License”.
For SUSE trademarks, see http://www.suse.com/company/legal/. All other third-party trademarks are the property of their respective owners. Trademark symbols (®, ™ etc.) denote trademarks of SUSE and its affiliates. Asterisks (*) denote third-party trademarks.
All information found in this book has been compiled with utmost attention to detail. However, this does not guarantee complete accuracy. Neither SUSE LLC, its affiliates, the authors nor the translators shall be held liable for possible errors or the consequences thereof.
This guide is intended for administrators who need to set up, configure, and maintain clusters with SUSE® Linux Enterprise High Availability Extension. For quick and efficient configuration and administration, the product includes both a graphical user interface and a command line interface (CLI). For performing key tasks, both approaches are covered in this guide. Thus, you can choose the appropriate tool that matches your needs.
This guide is divided into the following parts:
Before starting to install and configure your cluster, make yourself familiar with cluster fundamentals and architecture, get an overview of the key features and benefits. Learn which hardware and software requirements must be met and what preparations to take before executing the next steps. Perform the installation and basic setup of your HA cluster using YaST. Learn how to upgrade your cluster to the most recent release version or how to update individual packages.
Add, configure and manage cluster resources with either the Web interface (Hawk2), or the command line interface (crmsh). To avoid unauthorized access to the cluster configuration, define roles and assign them to certain users for fine-grained control. Learn how to use load balancing and fencing. If you consider writing your own resource agents or modifying existing ones, get some background information on how to create different types of resource agents.
SUSE Linux Enterprise High Availability Extension ships with the cluster-aware file systems OCFS2 and GFS2, and the Cluster Logical Volume Manager (Cluster LVM). For replication of your data, use DRBD*. It lets you mirror the data of a High Availability service from the active node of a cluster to its standby node. Furthermore, a clustered Samba server also provides a High Availability solution for heterogeneous environments.
Contains an overview of common problems and their solution. Presents the naming conventions used in this documentation with regard to clusters, resources and constraints. Contains a glossary with HA-specific terminology.
Documentation for our products is available at http://www.suse.com/documentation/, where you can also find the latest updates, and browse or download the documentation in various formats. The latest documentation updates can usually be found in the English language version.
The following documentation is available for this product:
This document guides you through the setup of a very basic two-node cluster,
using the bootstrap scripts provided by the
ha-cluster-bootstrap
package.
This includes the configuration of a virtual IP address as a cluster
resource and the use of SBD on shared storage as a node fencing mechanism.
This guide is intended for administrators who need to set up, configure, and maintain clusters with SUSE® Linux Enterprise High Availability Extension. For quick and efficient configuration and administration, the product includes both a graphical user interface and a command line interface (CLI). For performing key tasks, both approaches are covered in this guide. Thus, you can choose the appropriate tool that matches your needs.
This document describes how to set up highly available NFS storage in a two-node cluster, using the following components: DRBD* (Distributed Replicated Block Device), LVM (Logical Volume Manager), and Pacemaker as cluster resource manager.
This document guides you through the setup of a High Availability cluster with a remote
node or a guest node, managed by Pacemaker and pacemaker_remote
.
Remote in pacemaker_remote
does not refer to physical distance, but to the special status of nodes that
do not run the complete cluster stack and thus are not regular members of the
cluster.
Geo clustering protects workloads across globally distributed data
centers. This document guides you through the basic setup of a
Geo cluster, using the Geo bootstrap scripts provided by the
ha-cluster-bootstrap
package.
This document covers the setup options and parameters for Geo clusters and their components, such as booth ticket manager, the specific Csync2 setup, and the configuration of the required cluster resources (and how to transfer them to other sites in case of changes). Learn how to monitor and manage Geo clusters from command line or with the Hawk2 Web interface.
Your feedback and contribution to this documentation is welcome! Several channels are available:
For services and support options available for your product, refer to http://www.suse.com/support/.
To open a service request, you need a subscription at SUSE Customer Center. Go to https://scc.suse.com/support/requests, log in, and click .
Report issues with the documentation at https://bugzilla.suse.com/. To simplify this process, you can use the links next to headlines in the HTML version of this document. These preselect the right product and category in Bugzilla and add a link to the current section. You can start typing your bug report right away. A Bugzilla account is required.
To contribute to this documentation, use the
links next to headlines in the HTML version of this document. They take you to the source code on GitHub, where you can open a pull request. A GitHub account is required.Alternatively, you can report errors and send feedback concerning the documentation to <doc-team@suse.com>. Make sure to include the document title, the product version and the publication date of the documentation. Refer to the relevant section number and title (or include the URL) and provide a concise description of the problem.
The following notices and typographical conventions are used in this documentation:
tux >
command
Commands that can be run by any user, including the root
user.
root #
command
Commands that must be run with root
privileges. Often you
can also prefix these commands with the sudo
command to
run them.
crm(live)#
Commands executed in the interactive crm shell. For details, see Chapter 8, Configuring and Managing Cluster Resources (Command Line).
/etc/passwd
: directory names and file names
PLACEHOLDER: replace PLACEHOLDER with the actual value
PATH
: the environment variable PATH
ls
, --help
: commands, options, and
parameters
user
: users or groups
packagename: name of a package
Alt, Alt–F1: a key to press or a key combination; keys are shown in uppercase as on a keyboard
, › : menu items, buttons
amd64, em64t, ipf
This paragraph is only relevant for the architectures
amd64
, em64t
, and
ipf
. The arrows mark the beginning and the end of the
text block.
Dancing Penguins (Chapter Penguins, ↑Another Manual): This is a reference to a chapter in another manual.
Notices
Vital information you must be aware of before proceeding. Warns you about security issues, potential loss of data, damage to hardware, or physical hazards.
Important information you should be aware of before proceeding.
Additional information, for example about differences in software versions.
Helpful information, like a guideline or a piece of practical advice.
For an overview of naming conventions with regard to cluster nodes and names, resources, and constraints, see Appendix B, Naming Conventions.
This documentation is written in GeekoDoc, a subset of DocBook 5. The XML source files were validated by
jing
(see https://code.google.com/p/jing-trang/),
processed by xsltproc
, and converted into XSL-FO
using a customized version of Norman Walsh's stylesheets. The final
PDF is formatted through FOP from Apache Software Foundation. The open source
tools and the environment used to build this documentation are
provided by the DocBook Authoring and Publishing Suite (DAPS). The
project's home page can be found at https://github.com/openSUSE/daps.
The XML source code of this documentation can be found at https://github.com/SUSE/doc-sleha.
SUSE® Linux Enterprise High Availability Extension is an integrated suite of open source clustering technologies. It enables you to implement highly available physical and virtual Linux clusters, and to eliminate single points of failure. It ensures the high availability and manageability of critical network resources including data, applications, and services. Thus, it helps you maintain business continuity, protect data integrity, and reduce unplanned downtime for your mission-critical Linux workloads.
It ships with essential monitoring, messaging, and cluster resource management functionality (supporting failover, failback, and migration (load balancing) of individually managed cluster resources).
This chapter introduces the main product features and benefits of the High Availability Extension. Inside you will find several example clusters and learn about the components making up a cluster. The last section provides an overview of the architecture, describing the individual architecture layers and processes within the cluster.
For explanations of some common terms used in the context of High Availability clusters, refer to Glossary.
The following section informs you about system requirements, and some prerequisites for SUSE® Linux Enterprise High Availability Extension. It also includes recommendations for cluster setup.
If you are setting up a High Availability cluster with SUSE® Linux Enterprise High Availability Extension for the first time, the easiest way is to start with a basic two-node cluster. You can also use the two-node cluster to run some tests. Afterward, you can add more nodes by cloning existing cluster nodes with AutoYaST. The cloned nodes will have the same packages installed and the same system configuration as the original ones.
If you want to upgrade an existing cluster that runs an older version of SUSE Linux Enterprise High Availability Extension, refer to Chapter 5, Upgrading Your Cluster and Updating Software Packages.
The YaST cluster module allows you to set up a cluster manually (from scratch) or to modify options for an existing cluster.
However, if you prefer an automated approach for setting up a cluster,
refer to Article “Installation and Setup Quick Start”. It describes how to install the
needed packages and leads you to a basic two-node cluster, which is
set up with the ha-cluster-bootstrap
scripts.
You can also use a combination of both setup methods, for example: set up one node with YaST cluster and then use one of the bootstrap scripts to integrate more nodes (or vice versa).
This chapter covers two different scenarios: upgrading a cluster to another version of SUSE Linux Enterprise High Availability Extension (either a major release or a service pack) as opposed to updating individual packages on cluster nodes. See Section 5.2, “Upgrading your Cluster to the Latest Product Version” versus Section 5.3, “Updating Software Packages on Cluster Nodes”.
If you want to upgrade your cluster, check Section 5.2.1, “Supported Upgrade Paths for SLE HA and SLE HA Geo” and Section 5.2.2, “Required Preparations Before Upgrading” before starting to upgrade.
SUSE® Linux Enterprise High Availability Extension is an integrated suite of open source clustering technologies. It enables you to implement highly available physical and virtual Linux clusters, and to eliminate single points of failure. It ensures the high availability and manageability of critical network resources including data, applications, and services. Thus, it helps you maintain business continuity, protect data integrity, and reduce unplanned downtime for your mission-critical Linux workloads.
It ships with essential monitoring, messaging, and cluster resource management functionality (supporting failover, failback, and migration (load balancing) of individually managed cluster resources).
This chapter introduces the main product features and benefits of the High Availability Extension. Inside you will find several example clusters and learn about the components making up a cluster. The last section provides an overview of the architecture, describing the individual architecture layers and processes within the cluster.
For explanations of some common terms used in the context of High Availability clusters, refer to Glossary.
The High Availability Extension is available as an extension to SUSE Linux Enterprise Server 15 SP1. Support for using High Availability clusters across unlimited distances is available with Geo Clustering for SUSE Linux Enterprise High Availability Extension.
SUSE® Linux Enterprise High Availability Extension helps you ensure and manage the availability of your network resources. The following sections highlight some of the key features:
The High Availability Extension supports the following scenarios:
Active/active configurations
Active/passive configurations: N+1, N+M, N to 1, N to M
Hybrid physical and virtual clusters, allowing virtual servers to be clustered with physical servers. This improves service availability and resource usage.
Local clusters
Metro clusters (“stretched” local clusters)
Geo clusters (geographically dispersed clusters)
Your cluster can contain up to 32 Linux servers. Using pacemaker_remote, the cluster can be extended to include additional Linux servers beyond this limit. Any server in the cluster can restart resources (applications, services, IP addresses, and file systems) from a failed server in the cluster.
The High Availability Extension ships with Corosync messaging and membership layer and Pacemaker Cluster Resource Manager. Using Pacemaker, administrators can continually monitor the health and status of their resources, and manage dependencies. They can automatically stop and start services based on highly configurable rules and policies. The High Availability Extension allows you to tailor a cluster to the specific applications and hardware infrastructure that fit your organization. Time-dependent configuration enables services to automatically migrate back to repaired nodes at specified times.
With the High Availability Extension you can dynamically assign and reassign server storage as needed. It supports Fibre Channel or iSCSI storage area networks (SANs). Shared disk systems are also supported, but they are not a requirement. SUSE Linux Enterprise High Availability Extension also comes with a cluster-aware file system (OCFS2) and the cluster Logical Volume Manager (cluster LVM2). For replication of your data, use DRBD* to mirror the data of a High Availability service from the active node of a cluster to its standby node. Furthermore, SUSE Linux Enterprise High Availability Extension also supports CTDB (Cluster Trivial Database), a technology for Samba clustering.
SUSE Linux Enterprise High Availability Extension supports the mixed clustering of both physical and virtual Linux servers. SUSE Linux Enterprise Server 15 SP1 ships with Xen, an open source virtualization hypervisor, and with KVM (Kernel-based Virtual Machine). KVM is a virtualization software for Linux which is based on hardware virtualization extensions. The cluster resource manager in the High Availability Extension can recognize, monitor, and manage services running within virtual servers and services running in physical servers. Guest systems can be managed as services by the cluster.
SUSE Linux Enterprise High Availability Extension has been extended to support different geographical scenarios. Support for geographically dispersed clusters (Geo clusters) is available with Geo Clustering for SUSE Linux Enterprise High Availability Extension.
A single cluster in one location (for example, all nodes are located in one data center). The cluster uses multicast or unicast for communication between the nodes and manages failover internally. Network latency can be neglected. Storage is typically accessed synchronously by all nodes.
A single cluster that can stretch over multiple buildings or data centers, with all sites connected by fibre channel. The cluster uses multicast or unicast for communication between the nodes and manages failover internally. Network latency is usually low (<5 ms for distances of approximately 20 miles). Storage is frequently replicated (mirroring or synchronous replication).
Multiple, geographically dispersed sites with a local cluster each. The sites communicate via IP. Failover across the sites is coordinated by a higher-level entity. Geo clusters need to cope with limited network bandwidth and high latency. Storage is replicated asynchronously.
The greater the geographical distance between individual cluster nodes, the more factors may potentially disturb the high availability of services the cluster provides. Network latency, limited bandwidth and access to storage are the main challenges for long-distance clusters.
SUSE Linux Enterprise High Availability Extension includes a huge number of resource agents to manage
resources such as Apache, IPv4, IPv6 and many more. It also ships with
resource agents for popular third party applications such as IBM
WebSphere Application Server. For an overview of Open Cluster Framework
(OCF) resource agents included with your product, use the crm
ra
command as described in
Section 8.1.3, “Displaying Information about OCF Resource Agents”.
The High Availability Extension ships with a set of powerful tools. Use them for basic installation and setup of your cluster and for effective configuration and administration:
A graphical user interface for general system installation and administration. Use it to install the High Availability Extension on top of SUSE Linux Enterprise Server as described in the Installation and Setup Quick Start. YaST also provides the following modules in the High Availability category to help configure your cluster or individual components:
Cluster: Basic cluster setup. For details, refer to Chapter 4, Using the YaST Cluster Module.
DRBD: Configuration of a Distributed Replicated Block Device.
IP Load Balancing: Configuration of load balancing with Linux Virtual Server or HAProxy. For details, refer to Chapter 14, Load Balancing.
A user-friendly Web-based interface with which you can monitor and administer your High Availability clusters from Linux or non-Linux machines alike. Hawk2 can be accessed from any machine inside or outside of the cluster by using a (graphical) Web browser. Therefore it is the ideal solution even if the system on which you are working only provides a minimal graphical user interface. For details, Chapter 7, Configuring and Managing Cluster Resources with Hawk2.
crm
Shell
A powerful unified command line interface to configure resources and execute all monitoring or administration tasks. For details, refer to Chapter 8, Configuring and Managing Cluster Resources (Command Line).
The High Availability Extension allows you to configure up to 32 Linux servers into a high-availability cluster (HA cluster). Resources can be dynamically switched or moved to any node in the cluster. Resources can be configured to automatically migrate if a node fails, or they can be moved manually to troubleshoot hardware or balance the workload.
The High Availability Extension provides high availability from commodity components. Lower costs are obtained through the consolidation of applications and operations onto a cluster. The High Availability Extension also allows you to centrally manage the complete cluster. You can adjust resources to meet changing workload requirements (thus, manually “load balance” the cluster). Allowing clusters of more than two nodes also provides savings by allowing several nodes to share a “hot spare”.
An equally important benefit is the potential reduction of unplanned service outages and planned outages for software and hardware maintenance and upgrades.
Reasons that you would want to implement a cluster include:
Increased availability
Improved performance
Low cost of operation
Scalability
Disaster recovery
Data protection
Server consolidation
Storage consolidation
Shared disk fault tolerance can be obtained by implementing RAID on the shared disk subsystem.
The following scenario illustrates some benefits the High Availability Extension can provide.
Suppose you have configured a three-node cluster, with a Web server installed on each of the three nodes in the cluster. Each of the nodes in the cluster hosts two Web sites. All the data, graphics, and Web page content for each Web site are stored on a shared disk subsystem connected to each of the nodes in the cluster. The following figure depicts how this setup might look.
During normal cluster operation, each node is in constant communication with the other nodes in the cluster and performs periodic polling of all registered resources to detect failure.
Suppose Web Server 1 experiences hardware or software problems and the users depending on Web Server 1 for Internet access, e-mail, and information lose their connections. The following figure shows how resources are moved when Web Server 1 fails.
Web Site A moves to Web Server 2 and Web Site B moves to Web Server 3. IP addresses and certificates also move to Web Server 2 and Web Server 3.
When you configured the cluster, you decided where the Web sites hosted on each Web server would go should a failure occur. In the previous example, you configured Web Site A to move to Web Server 2 and Web Site B to move to Web Server 3. This way, the workload formerly handled by Web Server 1 continues to be available and is evenly distributed between any surviving cluster members.
When Web Server 1 failed, the High Availability Extension software did the following:
Detected a failure and verified with STONITH that Web Server 1 was really dead. STONITH is an acronym for “Shoot The Other Node In The Head”. It is a means of bringing down misbehaving nodes to prevent them from causing trouble in the cluster.
Remounted the shared data directories that were formerly mounted on Web server 1 on Web Server 2 and Web Server 3.
Restarted applications that were running on Web Server 1 on Web Server 2 and Web Server 3.
Transferred IP addresses to Web Server 2 and Web Server 3.
In this example, the failover process happened quickly and users regained access to Web site information within seconds, usually without needing to log in again.
Now suppose the problems with Web Server 1 are resolved, and Web Server 1 is returned to a normal operating state. Web Site A and Web Site B can either automatically fail back (move back) to Web Server 1, or they can stay where they are. This depends on how you configured the resources for them. Migrating the services back to Web Server 1 will incur some down-time. Therefore the High Availability Extension also allows you to defer the migration until a period when it will cause little or no service interruption. There are advantages and disadvantages to both alternatives.
The High Availability Extension also provides resource migration capabilities. You can move applications, Web sites, etc. to other servers in your cluster as required for system management.
For example, you could have manually moved Web Site A or Web Site B from Web Server 1 to either of the other servers in the cluster. Use cases for this are upgrading or performing scheduled maintenance on Web Server 1, or increasing performance or accessibility of the Web sites.
Cluster configurations with the High Availability Extension might or might not include a shared disk subsystem. The shared disk subsystem can be connected via high-speed Fibre Channel cards, cables, and switches, or it can be configured to use iSCSI. If a node fails, another designated node in the cluster automatically mounts the shared disk directories that were previously mounted on the failed node. This gives network users continuous access to the directories on the shared disk subsystem.
When using a shared disk subsystem with LVM2, that subsystem must be connected to all servers in the cluster from which it needs to be accessed.
Typical resources might include data, applications, and services. The following figures show how a typical Fibre Channel cluster configuration might look. The green lines depict connections to an Ethernet power switch. Such a device can be controlled over a network and can reboot a node when a ping request fails.
Although Fibre Channel provides the best performance, you can also configure your cluster to use iSCSI. iSCSI is an alternative to Fibre Channel that can be used to create a low-cost Storage Area Network (SAN). The following figure shows how a typical iSCSI cluster configuration might look.
Although most clusters include a shared disk subsystem, it is also possible to create a cluster without a shared disk subsystem. The following figure shows how a cluster without a shared disk subsystem might look.
This section provides a brief overview of the High Availability Extension architecture. It identifies and provides information on the architectural components, and describes how those components interoperate.
The High Availability Extension has a layered architecture. Figure 1.6, “Architecture” illustrates the different layers and their associated components.
This component provides reliable messaging, membership, and quorum information about the cluster. This is handled by the Corosync cluster engine, a group communication system.
Pacemaker as cluster resource manager is the “brain”
which reacts to events occurring in the cluster. It is implemented as
pacemaker-controld
, the cluster
controller, which coordinates all actions. Events can be nodes that join
or leave the cluster, failure of resources, or scheduled activities such
as maintenance, for example.
The local resource manager is located between the Pacemaker layer and the
resources layer on each node. It is implemented as pacemaker-execd
daemon. Through this daemon,
Pacemaker can start, stop, and monitor resources.
On every node, Pacemaker maintains the cluster information database
(CIB). It is an XML representation of the cluster configuration
(including cluster options, nodes, resources, constraints and the
relationship to each other). The CIB also reflects the current cluster
status. Each cluster node contains a CIB replica, which is synchronized
across the whole cluster. The pacemaker-based
daemon takes care of reading and writing cluster configuration and
status.
The DC is elected from all nodes in the cluster. This happens if there is no DC yet or if the current DC leaves the cluster for any reason. The DC is the only entity in the cluster that can decide that a cluster-wide change needs to be performed, such as fencing a node or moving resources around. All other nodes get their configuration and resource allocation information from the current DC.
The policy engine runs on every node, but the one on the DC is the active
one. The engine is implemented as
pacemaker-schedulerd
daemon.
When a cluster transition is needed, based on the current state and
configuration, pacemaker-schedulerd
calculates the expected next state of the cluster. It determines what
actions need to be scheduled to achieve the next state.
In a High Availability cluster, the services that need to be highly available are called resources. Resource agents (RAs) are scripts that start, stop, and monitor cluster resources.
The pacemakerd
daemon launches and
monitors all other related daemons. The daemon that coordinates all actions,
pacemaker-controld
, has an instance on
each cluster node. Pacemaker centralizes all cluster decision-making by
electing one of those instances as a master. Should the elected pacemaker-controld
daemon fail, a new one is
established.
Many actions performed in the cluster will cause a cluster-wide change. These actions can include things like adding or removing a cluster resource or changing resource constraints. It is important to understand what happens in the cluster when you perform such an action.
For example, suppose you want to add a cluster IP address resource. To do this, you can use the crm shell or the Web interface to modify the CIB. It is not required to perform the actions on the DC. You can use either tool on any node in the cluster and they will be relayed to the DC. The DC will then replicate the CIB change to all cluster nodes.
Based on the information in the CIB, the pacemaker-schedulerd
then computes the ideal
state of the cluster and how it should be achieved. It feeds a list of
instructions to the DC. The DC sends commands via the messaging/infrastructure
layer which are received by the pacemaker-controld
peers on
other nodes. Each of them uses its local resource agent executor (implemented
as pacemaker-execd
) to perform
resource modifications. The pacemaker-execd
is not cluster-aware and interacts
directly with resource agents.
All peer nodes report the results of their operations back to the DC.
After the DC concludes that all necessary operations are successfully
performed in the cluster, the cluster will go back to the idle state and
wait for further events. If any operation was not carried out as
planned, the pacemaker-schedulerd
is invoked again with the new information recorded in
the CIB.
In some cases, it may be necessary to power off nodes to protect shared
data or complete resource recovery. In a Pacemaker cluster, the implementation
of node level fencing is STONITH. For this, Pacemaker comes with a
fencing subsystem, pacemaker-fenced
.
STONITH devices have to be configured as cluster resources (that use
specific fencing agents), because this allows to monitor the fencing devices.
When clients detect a failure, they send a request to pacemaker-fenced
,
which then executes the fencing agent to bring down the node.
The following section informs you about system requirements, and some prerequisites for SUSE® Linux Enterprise High Availability Extension. It also includes recommendations for cluster setup.
The following list specifies hardware requirements for a cluster based on SUSE® Linux Enterprise High Availability Extension. These requirements represent the minimum hardware configuration. Additional hardware might be necessary, depending on how you intend to use your cluster.
1 to 32 Linux servers with software as specified in Section 2.2, “Software Requirements”.
The servers can be bare metal or virtual machines. They do not require identical hardware (memory, disk space, etc.), but they must have the same architecture. Cross-platform clusters are not supported.
Using pacemaker_remote
, the cluster can be
extended to include additional Linux servers beyond the 32-node limit.
At least two TCP/IP communication media per cluster node. The network equipment must support the communication means you want to use for cluster communication: multicast or unicast. The communication media should support a data rate of 100 Mbit/s or higher. For a supported cluster setup two or more redundant communication paths are required. This can be done via:
Network Device Bonding (preferred).
A second communication channel in Corosync.
For details, refer to Chapter 13, Network Device Bonding and Procedure 4.3, “Defining a Redundant Communication Channel”, respectively.
To avoid a “split brain” scenario, clusters need a node fencing mechanism. In a split brain scenario, cluster nodes are divided into two or more groups that do not know about each other (because of a hardware or software failure or because of a cut network connection). A fencing mechanism isolates the node in question (usually by resetting or powering off the node). This is also called STONITH (“Shoot the other node in the head”). A node fencing mechanism can be either a physical device (a power switch) or a mechanism like SBD (STONITH by disk) in combination with a watchdog. Using SBD requires shared storage.
Unless SBD is used, each node in the High Availability cluster must have at least one STONITH device. We strongly recommend multiple STONITH devices per node.
You must have a node fencing mechanism for your cluster.
The global cluster options
stonith-enabled
and
startup-fencing
must be set to
true
.
When you change them, you lose support.
All nodes that will be part of the cluster need at least the following modules and extensions:
Base System Module 15 SP1
Server Applications Module 15 SP1
SUSE Linux Enterprise High Availability Extension 15 SP1
Depending on the system roles
you select during
installation, the following software patterns are installed by default:
System Role |
Software Pattern (YaST/Zypper) |
---|---|
HA Node |
|
HA GEO Node |
|
An installation via those system roles results in a minimal installation only. You might need to add more packages manually, if required.
For machines that originally had another system role assigned, you need to
manually install the sles_ha
or
ha_geo
patterns and any further packages that you
need.
Some services require shared storage. If using an external NFS share, it must be reliably accessible from all cluster nodes via redundant communication paths.
To make data highly available, a shared disk system (Storage Area Network, or SAN) is recommended for your cluster. If a shared disk subsystem is used, ensure the following:
The shared disk system is properly set up and functional according to the manufacturer’s instructions.
The disks contained in the shared disk system should be configured to use mirroring or RAID to add fault tolerance to the shared disk system.
If you are using iSCSI for shared disk system access, ensure that you have properly configured iSCSI initiators and targets.
When using DRBD* to implement a mirroring RAID system that distributes data across two machines, make sure to only access the device provided by DRBD—never the backing device. To leverage the redundancy it is possible to use the same NICs as the rest of the cluster.
When using SBD as STONITH mechanism, additional requirements apply for the shared storage. For details, see Section 11.3, “Requirements”.
For a supported and useful High Availability setup, consider the following recommendations:
For clusters with more than two nodes, it is strongly recommended to use an odd number of cluster nodes to have quorum. For more information about quorum, see Section 6.2, “Quorum Determination”.
Cluster nodes must synchronize to an NTP server outside the cluster. Since SUSE Linux Enterprise High Availability Extension 15, chrony is the default implementation of NTP. For more information, see the Administration Guide for SUSE Linux Enterprise Server 15 SP1, chapter Time Synchronization with NTP. It is available from http://www.suse.com/documentation/.
If nodes are not synchronized, the cluster may not work properly. In addition, log files and cluster reports are very hard to analyze without synchronization. If you use the bootstrap scripts, you will be warned if NTP is not configured yet.
Must be identical on all nodes.
Use static IP addresses.
List all cluster nodes in the /etc/hosts
file
with their fully qualified host name and short host name. It is essential that
members of the cluster can find each other by name. If the names are not
available, internal cluster communication will fail.
For details on how Pacemaker gets the node names, see also http://clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/s-node-name.html.
All cluster nodes must be able to access each other via SSH. Tools
like crm report
(for troubleshooting) and
Hawk2's require passwordless
SSH access between the nodes,
otherwise they can only collect data from the current node.
If passwordless SSH access does not comply with regulatory
requirements, you can use the work-around described in
Appendix D, Running Cluster Reports Without root
Access for running
crm report
.
For the
there is currently no alternative for passwordless login.If you are setting up a High Availability cluster with SUSE® Linux Enterprise High Availability Extension for the first time, the easiest way is to start with a basic two-node cluster. You can also use the two-node cluster to run some tests. Afterward, you can add more nodes by cloning existing cluster nodes with AutoYaST. The cloned nodes will have the same packages installed and the same system configuration as the original ones.
If you want to upgrade an existing cluster that runs an older version of SUSE Linux Enterprise High Availability Extension, refer to Chapter 5, Upgrading Your Cluster and Updating Software Packages.
For the manual installation of the packages for High Availability Extension refer to Article “Installation and Setup Quick Start”. It leads you through the setup of a basic two-node cluster.
After you have installed and set up a two-node cluster, you can extend the cluster by cloning existing nodes with AutoYaST and adding the clones to the cluster.
AutoYaST uses profiles that contains installation and configuration data. A profile tells AutoYaST what to install and how to configure the installed system to get a ready-to-use system in the end. This profile can then be used for mass deployment in different ways (for example, to clone existing cluster nodes).
For detailed instructions on how to use AutoYaST in various scenarios, see the SUSE Linux Enterprise 15 SP1 AutoYaST Guide, available from http://www.suse.com/documentation/.
Procedure 3.1, “Cloning a Cluster Node with AutoYaST” assumes you are rolling out SUSE Linux Enterprise High Availability Extension 15 SP1 to a set of machines with identical hardware configurations.
If you need to deploy cluster nodes on non-identical hardware, refer to the SUSE Linux Enterprise 15 SP1 Deployment Guide, chapter Automated Installation, section Rule-Based Autoinstallation.
Make sure the node you want to clone is correctly installed and configured. For details, see the Installation and Setup Quick Start for SUSE Linux Enterprise High Availability Extension or Chapter 4, Using the YaST Cluster Module.
Follow the description outlined in the SUSE Linux Enterprise 15 SP1 Deployment Guide for simple mass installation. This includes the following basic steps:
Creating an AutoYaST profile. Use the AutoYaST GUI to create and modify a profile based on the existing system configuration. In AutoYaST, choose the
module and click the button. If needed, adjust the configuration in the other modules and save the resulting control file as XML.If you have configured DRBD, you can select and clone this module in the AutoYaST GUI, too.
Determining the source of the AutoYaST profile and the parameter to pass to the installation routines for the other nodes.
Determining the source of the SUSE Linux Enterprise Server and SUSE Linux Enterprise High Availability Extension installation data.
Determining and setting up the boot scenario for autoinstallation.
Passing the command line to the installation routines, either by
adding the parameters manually or by creating an
info
file.
Starting and monitoring the autoinstallation process.
After the clone has been successfully installed, execute the following steps to make the cloned node join the cluster:
Transfer the key configuration files from the already configured nodes to the cloned node with Csync2 as described in Section 4.5, “Transferring the Configuration to All Nodes”.
To bring the node online, start the Pacemaker service on the cloned node as described in Section 4.8, “Bringing the Cluster Online”.
The cloned node will now join the cluster because the
/etc/corosync/corosync.conf
file has been applied to
the cloned node via Csync2. The CIB is automatically synchronized
among the cluster nodes.
The YaST cluster module allows you to set up a cluster manually (from scratch) or to modify options for an existing cluster.
However, if you prefer an automated approach for setting up a cluster,
refer to Article “Installation and Setup Quick Start”. It describes how to install the
needed packages and leads you to a basic two-node cluster, which is
set up with the ha-cluster-bootstrap
scripts.
You can also use a combination of both setup methods, for example: set up one node with YaST cluster and then use one of the bootstrap scripts to integrate more nodes (or vice versa).
Several key terms used in the YaST cluster module and in this chapter are defined below.
bindnetaddr
)
The network address the Corosync executive should bind to. To simplify sharing configuration files across
the cluster, Corosync uses network interface netmask to mask only
the address bits that are used for routing the network. For example,
if the local interface is 192.168.5.92
with netmask
255.255.255.0
, set
bindnetaddr
to
192.168.5.0
. If the local interface is
192.168.5.92
with netmask
255.255.255.192
, set
bindnetaddr
to
192.168.5.64
.
As the same Corosync configuration will be used on all nodes,
make sure to use a network address as
bindnetaddr
, not the address of a specific
network interface.
conntrack
ToolsAllow interaction with the in-kernel connection tracking system for enabling stateful packet inspection for iptables. Used by the High Availability Extension to synchronize the connection status between cluster nodes. For detailed information, refer to http://conntrack-tools.netfilter.org/.
A synchronization tool that can be used to replicate configuration files
across all nodes in the cluster, and even across Geo clusters. Csync2 can handle any number of hosts, sorted into
synchronization groups. Each synchronization group has its own list of
member hosts and its include/exclude patterns that define which files
should be synchronized in the synchronization group. The groups, the
host names belonging to each group, and the include/exclude rules for
each group are specified in the Csync2 configuration file,
/etc/csync2/csync2.cfg
.
For authentication, Csync2 uses the IP addresses and pre-shared keys within a synchronization group. You need to generate one key file for each synchronization group and copy it to all group members.
For more information about Csync2, refer to http://oss.linbit.com/csync2/paper.pdf
The term “existing cluster” is used to refer to any cluster that consists of at least one node. Existing clusters have a basic Corosync configuration that defines the communication channels, but they do not necessarily have resource configuration yet.
A technology used for a one-to-many communication within a network that can be used for cluster communication. Corosync supports both multicast and unicast.
To use multicast for cluster communication, make sure your switches support multicast.
mcastaddr
)
IP address to be used for multicasting by the Corosync executive. The IP address can either be IPv4 or IPv6. If IPv6 networking is used, node IDs must be specified. You can use any multicast address in your private network.
mcastport
)
The port to use for cluster communication. Corosync uses two ports: the specified
mcastport
for receiving multicast, and
mcastport -1
for sending multicast.
Allows the use of multiple redundant local area networks for resilience against partial or total network faults. This way, cluster communication can still be kept up as long as a single network is operational. Corosync supports the Totem Redundant Ring Protocol. A logical token-passing ring is imposed on all participating nodes to deliver messages in a reliable and sorted manner. A node is allowed to broadcast a message only if it holds the token.
When having defined redundant communication channels in Corosync,
use RRP to tell the cluster how to use these interfaces. RRP can have
three modes (rrp_mode
):
If set to active
, Corosync uses both
interfaces actively. However, this mode is deprecated.
If set to passive
, Corosync sends messages
alternatively over the available networks.
If set to none
, RRP is disabled.
A technology for sending messages to a single network destination. Corosync supports both multicast and unicast. In Corosync, unicast is implemented as UDP-unicast (UDPU).
Start YaST and select
› . Alternatively, start the module from command line:sudo yast2 cluster
The following list shows an overview of the available screens in the YaST cluster module. It also mentions whether the screen contains parameters that are required for successful cluster setup or whether its parameters are optional.
Allows you to define one or two communication channels for communication between the cluster nodes. As transport protocol, either use multicast (UDP) or unicast (UDPU). For details, see Section 4.3, “Defining the Communication Channels”.
For a supported cluster setup two or more redundant communication paths are required. The preferred way is to use network device bonding as described in Chapter 13, Network Device Bonding.
If this is impossible, you need to define a second communication channel in Corosync.
Allows you to define the authentication settings for the cluster. HMAC/SHA1 authentication requires a shared secret used to protect and authenticate messages. For details, see Section 4.4, “Defining Authentication Settings”.
Csync2 helps you to keep track of configuration changes and to keep files synchronized across the cluster nodes. For details, see Section 4.5, “Transferring the Configuration to All Nodes”.
Allows you to configure the user space
conntrackd
. Use the conntrack
tools for stateful packet inspection for iptables.
For details, see Section 4.6, “Synchronizing Connection Status Between Cluster Nodes”.
Allows you to configure the service for bringing the cluster node online. Define whether to start the Pacemaker service at boot time and whether to open the ports in the firewall that are needed for communication between the nodes. For details, see Section 4.7, “Configuring Services”.
If you start the cluster module for the first time, it appears as a wizard, guiding you through all the steps necessary for basic setup. Otherwise, click the categories on the left panel to access the configuration options for each step.
Some settings in the YaST cluster module apply only to the current node. Other settings may automatically be transferred to all nodes with Csync2. Find detailed information about this in the following sections.
For successful communication between the cluster nodes, define at least one communication channel. As transport protocol, either use multicast (UDP) or unicast (UDPU) as described in Procedure 4.1 or Procedure 4.2, respectively. If you want to define a second, redundant channel (Procedure 4.3), both communication channels must use the same protocol.
For deploying SUSE Linux Enterprise High Availability Extension in public cloud platforms, use unicast as transport protocol. Multicast is generally not supported by the cloud platforms themselves.
All settings defined in the YaST
/etc/corosync/corosync.conf
. Find example
files for a multicast and a unicast setup in
/usr/share/doc/packages/corosync/
.
If you are using IPv4 addresses, node IDs are optional. If you are using IPv6 addresses, node IDs are required. Instead of specifying IDs manually for each node, the YaST cluster module contains an option to automatically generate a unique ID for every cluster node.
When using multicast, the same bindnetaddr
,
mcastaddr
, and mcastport
will be used for all cluster nodes. All nodes in the cluster will know each
other by using the same multicast address. For different clusters, use
different multicast addresses.
Start the YaST cluster module and switch to the
category.
Set the Multicast
.
Define the
. Set the value to the subnet you will use for cluster multicast.Define the
.Define the
.To automatically generate a unique ID for every cluster node keep
enabled.Define a
.
Enter the number of quorum in case of a partitioned cluster. By
default, each node has 1
vote. The number of
must match the number of nodes in
your cluster.
Confirm your changes.
If needed, define a redundant communication channel in Corosync as described in Procedure 4.3, “Defining a Redundant Communication Channel”.
If you want to use unicast instead of multicast for cluster communication, proceed as follows.
Start the YaST cluster module and switch to the
category.
Set the Unicast
.
Define the
.For unicast communication, Corosync needs to know the IP addresses of all nodes in the cluster. For each node that will be part of the cluster, click
and enter the following details:
(only required if you use a second communication channel in Corosync)
(only required if the option is disabled)
To modify or remove any addresses of cluster members, use the
or buttons.To automatically generate a unique ID for every cluster node keep
enabled.Define a
.
Enter the number of quorum in case of a partitioned cluster. By
default, each node has 1
vote. The number of
must match the number of nodes in
your cluster.
Confirm your changes.
If needed, define a redundant communication channel in Corosync as described in Procedure 4.3, “Defining a Redundant Communication Channel”.
If network device bonding cannot be used for any reason, the second best choice is to define a redundant communication channel (a second ring) in Corosync. That way, two physically separate networks can be used for communication. If one network fails, the cluster nodes can still communicate via the other network.
The additional communication channel in
Corosync will form a second token-passing ring. In
/etc/corosync/corosync.conf
, the first channel you
configured is the primary ring and gets the ring number
0
. The second ring (redundant channel) gets the ring number
1
.
When having defined redundant communication channels in Corosync, use RRP to tell the cluster how to use these interfaces. With RRP, two physically separate networks are used for communication. If one network fails, the cluster nodes can still communicate via the other network.
RRP can have three modes:
If set to active
, Corosync uses both
interfaces actively. However, this mode is deprecated.
If set to passive
, Corosync sends messages
alternatively over the available networks.
If set to none
, RRP is disabled.
/etc/hosts
If multiple rings are configured in Corosync, each node can
have multiple IP addresses. This needs to be reflected in the
/etc/hosts
file of all nodes.
Start the YaST cluster module and switch to the
category.Activate
. The redundant channel must use the same protocol as the first communication channel you defined.If you use multicast, enter the following parameters: the
to use, the and the for the redundant channel.If you use unicast, define the following parameters: the
to use, and the . Enter the IP addresses of all nodes that will be part of the cluster.To tell Corosync how and when to use the different channels, select the
to use: If only one communication channel is defined,
none
).
If set to active
, Corosync uses both
interfaces actively. However, this mode is deprecated.
If set to passive
, Corosync sends messages
alternatively over the available networks.
When RRP is used, the High Availability Extension monitors the status of the current rings and automatically re-enables redundant rings after faults.
Alternatively, check the ring status manually with
corosync-cfgtool
. View the available options with
-h
.
Confirm your changes.
To define the authentication settings for the cluster, you can use HMAC/SHA1 authentication. This requires a shared secret used to protect and authenticate messages. The authentication key (password) you specify will be used on all nodes in the cluster.
Start the YaST cluster module and switch to the
category.Activate
. For a newly created cluster, click /etc/corosync/authkey
.
If you want the current machine to join an existing cluster, do not
generate a new key file. Instead, copy the
/etc/corosync/authkey
from one of the nodes to the
current machine (either manually or with Csync2).
Confirm your changes. YaST writes the configuration to
/etc/corosync/corosync.conf
.
Instead of copying the resulting configuration files to all nodes
manually, use the csync2
tool for replication across
all nodes in the cluster.
This requires the following basic steps:
Csync2 helps you to keep track of configuration changes and to keep files synchronized across the cluster nodes:
You can define a list of files that are important for operation.
You can show changes to these files (against the other cluster nodes).
You can synchronize the configured files with a single command.
With a simple shell script in ~/.bash_logout
, you
can be reminded about unsynchronized changes before logging out of the
system.
Find detailed information about Csync2 at http://oss.linbit.com/csync2/ and http://oss.linbit.com/csync2/paper.pdf.
Start the YaST cluster module and switch to the
category. To specify the synchronization group, click hostname
command.
If host name resolution does not work properly in your
network, you can also specify a combination of host name and IP address
for each cluster node. To do so, use the string
HOSTNAME@IP such as
alice@192.168.2.100
, for example. Csync2
will then use the IP addresses when connecting.
Click /etc/csync2/key_hagroup
. After it has been created,
it must be copied manually to all members of the cluster.
To populate the
list with the files that usually need to be synchronized among all nodes, click .To
, or files from the list of files to be synchronized use the respective buttons. You must enter the absolute path name for each file.Activate Csync2 by clicking
. This will execute the following command to start Csync2 automatically at boot time:root #
systemctl
enable csync2.socket
Confirm your changes. YaST writes the Csync2
configuration to /etc/csync2/csync2.cfg
.
To start the synchronization process now, proceed with Section 4.5.2, “Synchronizing Changes with Csync2”.
To successfully synchronize the files with Csync2, the following requirements must be met:
The same Csync2 configuration is available on all cluster nodes.
The same Csync2 authentication key is available on all cluster nodes.
Csync2 must be running on all cluster nodes.
Before the first Csync2 run, you therefore need to make the following preparations:
Copy the file /etc/csync2/csync2.cfg
manually to all nodes after you have configured it as described in Section 4.5.1, “Configuring Csync2 with YaST”.
Copy the file /etc/csync2/key_hagroup
that you
have generated on one node in Step 3
of Section 4.5.1
to all nodes in the cluster. It is needed for
authentication by Csync2. However, do not
regenerate the file on the other nodes—it needs to be the same
file on all nodes.
Execute the following command on all nodes to start the service now:
root #
systemctl
start csync2.socket
To initially synchronize all files once, execute the following command on the machine that you want to copy the configuration from:
root #
csync2
-xv
This will synchronize all the files once by pushing them to the other nodes. If all files are synchronized successfully, Csync2 will finish with no errors.
If one or several files that are to be synchronized have been modified on other nodes (not only on the current one), Csync2 reports a conflict. You will get an output similar to the one below:
While syncing file /etc/corosync/corosync.conf: ERROR from peer hex-14: File is also marked dirty here! Finished with 1 errors.
If you are sure that the file version on the current node is the “best” one, you can resolve the conflict by forcing this file and resynchronizing:
root #
csync2
-f
/etc/corosync/corosync.confroot #
csync2
-x
For more information on the Csync2 options, run
csync2 -help
Csync2 only pushes changes. It does not continuously synchronize files between the machines.
Each time you update files that need to be synchronized, you need to
push the changes to the other machines: Run csync2
-xv
on the machine where you did the changes. If you run
the command on any of the other machines with unchanged files, nothing will
happen.
To enable stateful packet inspection for iptables, configure and use the conntrack tools. This requires the following basic steps:
Configuring a resource for
conntrackd
(class:
ocf
, provider: heartbeat
). If
you use Hawk2 to add the resource, use the default values proposed
by Hawk2.
After having configured the conntrack tools, you can use them for Linux Virtual Server, see Load Balancing.
conntrackd
with YaST #
Use the YaST cluster module to configure the user space
conntrackd
. It needs a
dedicated network interface that is not used for other communication
channels. The daemon can be started via a resource agent afterward.
Start the YaST cluster module and switch to the
category.Select a
for synchronizing the connection status. The IPv4 address of the selected interface is automatically detected and shown in YaST. It must already be configured and it must support multicast.Define the
to be used for synchronizing the connection status.In
, define a numeric ID for the group to synchronize the connection status to.
Click conntrackd
.
If you modified any options for an existing cluster, confirm your changes and close the cluster module.
For further cluster configuration, click Section 4.7, “Configuring Services”.
and proceed withconntrackd
#In the YaST cluster module define whether to start certain services on a node at boot time. You can also use the module to start and stop the services manually. To bring the cluster nodes online and start the cluster resource manager, Pacemaker must be running as a service.
In the YaST cluster module, switch to the
category.To start Pacemaker each time this cluster node is booted, select the respective option in the
group. If you select in the group, you must start Pacemaker manually each time this node is booted. To start Pacemaker manually, use the command:root #
crm
cluster start
To start or stop Pacemaker immediately, click the respective button.
To open the ports in the firewall that are needed for cluster communication on the current machine, activate
.Confirm your changes. Note that the configuration only applies to the current machine, not to all cluster nodes.
After the initial cluster configuration is done, start the Pacemaker service on each cluster node to bring the stack online:
Log in to an existing node.
Check if the service is already running:
root #
crm
cluster status
If not, start Pacemaker now:
root #
crm
cluster start
Repeat the steps above for each of the cluster nodes.
On one of the nodes, check the cluster status with the
crm status
command. If all nodes are
online, the output should be similar to the following:
root #
crm status
Last updated: Thu Jul 3 11:07:10 2014
Last change: Thu Jul 3 10:58:43 2014
Current DC: alice (175704363) - partition with quorum
2 Nodes configured
0 Resources configured
Online: [ alice bob ]
This output indicates that the cluster resource manager is started and is ready to manage resources.
After the basic configuration is done and the nodes are online, you can start to configure cluster resources. Use one of the cluster management tools like the crm shell (crmsh) or Hawk2. For more information, see Chapter 8, Configuring and Managing Cluster Resources (Command Line) or Chapter 7, Configuring and Managing Cluster Resources with Hawk2.
This chapter covers two different scenarios: upgrading a cluster to another version of SUSE Linux Enterprise High Availability Extension (either a major release or a service pack) as opposed to updating individual packages on cluster nodes. See Section 5.2, “Upgrading your Cluster to the Latest Product Version” versus Section 5.3, “Updating Software Packages on Cluster Nodes”.
If you want to upgrade your cluster, check Section 5.2.1, “Supported Upgrade Paths for SLE HA and SLE HA Geo” and Section 5.2.2, “Required Preparations Before Upgrading” before starting to upgrade.
In the following, find definitions of the most important terms used in this chapter:
A major release is a new product version that brings new features and tools, and decommissions previously deprecated components. It comes with backward incompatible changes.
If a new product version includes major changes that are backward incompatible, the cluster needs to be upgraded by an offline migration. You need to take all nodes offline and upgrade the cluster as a whole, before you can bring all nodes back online.
In a rolling upgrade one cluster node at a time is upgraded while the rest of the cluster is still running. You take the first node offline, upgrade it and bring it back online to join the cluster. Then you continue one by one until all cluster nodes are upgraded to a major version.
Combines several patches into a form that is easy to install or deploy. Service packs are numbered and usually contain security fixes, updates, upgrades, or enhancements of programs.
Installation of a newer minor version of a package, which usually contains security fixes and other important fixes.
Installation of a newer major version of a package or distribution, which brings new features. See also Offline Migration versus Rolling Upgrade.
Which upgrade path is supported, and how to perform the upgrade depends on the current product version and on the target version you want to migrate to.
Rolling upgrades are only supported within the same major release (from the GA of a product version to the next service pack, and from one service pack to the next).
Offline migrations are required to upgrade from one major release to the next (for example, from SLE HA 12 to SLE HA 15) or from a service pack within one major release to the next major release (for example, from SLE HA 12 SP3 to SLE HA 15).
Section 5.2.1 gives an overview of the supported upgrade paths for SLE HA (Geo). The column For Details lists the specific upgrade documentation you should refer (including also the base system and Geo Clustering for SUSE Linux Enterprise High Availability Extension). This documentation is available from:
Mixed clusters running on SUSE Linux Enterprise High Availability Extension 12/SUSE Linux Enterprise High Availability Extension 15 are not supported.
After the upgrade process to product version 15, reverting back to product version 12 is not supported.
Upgrade From ... To |
Upgrade Path |
For Details |
---|---|---|
SLE HA 11 SP3 to SLE HA (Geo) 12 |
Offline Migration |
|
SLE HA (Geo) 11 SP4 to SLE HA (Geo) 12 SP1 |
Offline Migration |
|
SLE HA (Geo) 12 to SLE HA (Geo) 12 SP1 |
Rolling Upgrade |
|
SLE HA (Geo) 12 SP1 to SLE HA (Geo) 12 SP2 |
Rolling Upgrade |
|
SLE HA (Geo) 12 SP2 to SLE HA (Geo) 12 SP3 |
Rolling Upgrade |
|
SLE HA (Geo) 12 SP3 to SLE HA (Geo) 12 SP4 |
Rolling Upgrade |
|
SLE HA (Geo) 12 SP3 to SLE HA (Geo) 15 |
Offline Migration |
|
SLE HA (Geo) 12 SP4 to SLE HA (Geo) 15 SP1 |
Offline Migration |
|
SLE HA (Geo) 15 to SLE HA (Geo) 15 SP1 |
Rolling Upgrade |
|
Ensure that your system backup is up to date and restorable.
Test the upgrade procedure on a staging instance of your cluster setup first, before performing it in a production environment. This gives you an estimation of the time frame required for the maintenance window. It also helps to detect and solve any unexpected problems that might arise.
This section applies to the following scenarios:
Upgrading from SLE HA 11 SP3 to SLE HA 12—for details see Procedure 5.1, “Upgrading from Product Version 11 to 12: Cluster-Wide Offline Migration”.
Upgrading from SLE HA 11 SP4 to SLE HA 12 SP1—for details see Procedure 5.1, “Upgrading from Product Version 11 to 12: Cluster-Wide Offline Migration”.
Upgrading from SLE HA 12 SP3 to SLE HA 15—for details see Procedure 5.2, “Upgrading from Product Version 12 to 15: Cluster-Wide Offline Migration”.
Upgrading from SLE HA 12 SP4 to SLE HA 15 SP1—for details see Procedure 5.2, “Upgrading from Product Version 12 to 15: Cluster-Wide Offline Migration”.
If your cluster is still based on an older product version than the ones listed above, first upgrade it to a version of SLES and SLE HA that can be used as a source for upgrading to the desired target version.
The High Availability Extension 12 cluster stack comes with major changes in various
components (for example, /etc/corosync/corosync.conf
, disk formats of OCFS2).
Therefore, a rolling upgrade
from any SUSE Linux Enterprise High Availability Extension
11 version is not supported. Instead, all cluster nodes must be offline
and the cluster needs to be migrated as a whole as described in
Procedure 5.1, “Upgrading from Product Version 11 to 12: Cluster-Wide Offline Migration”.
Log in to each cluster node and stop the cluster stack with:
root #
rcopenais
stop
For each cluster node, perform an upgrade to the desired target version of SUSE Linux Enterprise Server and SUSE Linux Enterprise High Availability Extension—see Section 5.2.1, “Supported Upgrade Paths for SLE HA and SLE HA Geo”.
After the upgrade process has finished, reboot each node with the upgraded version of SUSE Linux Enterprise Server and SUSE Linux Enterprise High Availability Extension.
If you use OCFS2 in your cluster setup, update the on-device structure by executing the following command:
root #
o2cluster
--update PATH_TO_DEVICE
It adds additional parameters to the disk. They are needed for the updated OCFS2 version that is shipped with SUSE Linux Enterprise High Availability Extension 12 and 12 SPx.
To update /etc/corosync/corosync.conf
for Corosync version 2:
Log in to one node and start the YaST cluster module.
Switch to the Procedure 4.1, “Defining the First Communication Channel (Multicast)” or Procedure 4.2, “Defining the First Communication Channel (Unicast)”, respectively.
category and enter values for the following new parameters: and . For details, seeIf YaST should detect any other options that are invalid or missing according to Corosync version 2, it will prompt you to change them.
Confirm your changes in YaST. YaST will write them to
/etc/corosync/corosync.conf
.
If Csync2 is configured for your cluster, use the following command to push the updated Corosync configuration to the other cluster nodes:
root #
csync2
-xv
For details on Csync2, see Section 4.5, “Transferring the Configuration to All Nodes”.
Alternatively, synchronize the updated Corosync configuration by
manually copying /etc/corosync/corosync.conf
to all cluster nodes.
Log in to each node and start the cluster stack with:
root #
crm
cluster start
Check the cluster status with crm status
or with
Hawk2.
Configure the following services to start at boot time:
root #
systemctl enable pacemakerroot #
systemctl enable hawkroot #
systemctl enable sbd
Sometimes new features are only available with the latest CIB syntax version. When you upgrade to a new product version, your CIB syntax version will not be upgraded by default.
Check your version with:
cibadmin -Q | grep validate-with
Upgrade to the latest CIB syntax version with:
root #
cibadmin
--upgrade --force
If you decide to install the cluster nodes from scratch (instead of upgrading them), see Section 2.2, “Software Requirements” for the list of modules required for SUSE Linux Enterprise High Availability Extension 15 SP1. Find more information about modules, extensions and related products in the release notes for SUSE Linux Enterprise Server 15. They are available at https://www.suse.com/releasenotes/.
Before starting the offline migration to SUSE Linux Enterprise High Availability Extension 15, manually upgrade the CIB syntax in your current cluster as described in Note: Upgrading the CIB Syntax Version.
Log in to each cluster node and stop the cluster stack with:
root #
crm
cluster stop
For each cluster node, perform an upgrade to the desired target version of SUSE Linux Enterprise Server and SUSE Linux Enterprise High Availability Extension—see Section 5.2.1, “Supported Upgrade Paths for SLE HA and SLE HA Geo”.
After the upgrade process has finished, log in to each node and boot it with the upgraded version of SUSE Linux Enterprise Server and SUSE Linux Enterprise High Availability Extension.
If you use Cluster LVM, you need to migrate from clvmd to lvmlockd.
See the man page of lvmlockd
, section
changing a clvm VG to a lockd VG and Section 21.4, “Online Migration from Mirror LV to Cluster MD”.
Start the cluster stack with:
root #
crm
cluster start
Check the cluster status with crm status
or with
Hawk2.
This section applies to the following scenarios:
Upgrading from SLE HA 12 to SLE HA 12 SP1
Upgrading from SLE HA 12 SP1 to SLE HA 12 SP2
Upgrading from SLE HA 12 SP2 to SLE HA 12 SP3
Upgrading from SLE HA 15 to SLE HA 15 SP1
Before starting an upgrade for a node, stop the cluster stack on that node.
If the cluster resource manager on a node is active during the software update, this can lead to unpredictable results like fencing of active nodes.
Log in as root
on the node that you want to upgrade and stop the
cluster stack:
root #
crm
cluster stop
Perform an upgrade to the desired target version of SUSE Linux Enterprise Server and SUSE Linux Enterprise High Availability Extension. To find the details for the individual upgrade processes, see Section 5.2.1, “Supported Upgrade Paths for SLE HA and SLE HA Geo”.
Start the cluster stack on the upgraded node to make the node rejoin the cluster:
root #
crm
cluster start
Take the next node offline and repeat the procedure for that node.
Check the cluster status with crm status
or with
Hawk2.
The new features shipped with the latest product version will only be available after all cluster nodes have been upgraded to the latest product version. Mixed version clusters are only supported for a short time frame during the rolling upgrade. Complete the rolling upgrade within one week.
The Hawk2
screen also shows a warning if different CRM versions are detected for your cluster nodes.Before starting an update for a node, either stop the cluster stack on that node or put the node into maintenance mode, depending on whether the cluster stack is affected or not. See Step 1 for details.
If the cluster resource manager on a node is active during the software update, this can lead to unpredictable results like fencing of active nodes.
Before installing any package updates on a node, check the following:
Does the update affect any packages belonging to SUSE Linux Enterprise High Availability Extension or Geo Clustering for SUSE Linux Enterprise High Availability Extension?
If yes
: Stop the cluster stack on
the node before starting the software update:
root #
crm
cluster stop
Does the package update require a reboot? If yes
:
Stop the cluster stack on the node before starting the software
update:
root #
crm
cluster stop
If none of the situations above apply, you do not need to stop the cluster stack. In that case, put the node into maintenance mode before starting the software update:
root #
crm
node maintenance NODE_NAME
For more details on maintenance mode, see Section 16.2, “Different Options for Maintenance Tasks”.
Install the package update using either YaST or Zypper.
After the update has been successfully installed:
Either start the cluster stack on the respective node (if you stopped it in Step 1):
root #
crm
cluster start
or remove the maintenance flag to bring the node back to normal mode:
root #
crm
node ready NODE_NAME
Check the cluster status with crm status
or with
Hawk2.
For detailed information about any changes and new features of the product you are upgrading to, refer to its release notes. They are available from https://www.suse.com/releasenotes/.
The main purpose of an HA cluster is to manage user services. Typical examples of user services are an Apache Web server or a database. From the user's point of view, the services do something specific when ordered to do so. To the cluster, however, they are only resources which may be started or stopped—the nature of the service is irrelevant to the cluster.
In this chapter, we will introduce some basic concepts you need to know when configuring resources and administering your cluster. The following chapters show you how to execute the main configuration and administration tasks with each of the management tools the High Availability Extension provides.
To configure and manage cluster resources, either use Hawk2, or the crm shell (crmsh) command line utility. If you upgrade from an earlier version of SUSE® Linux Enterprise High Availability Extension where Hawk was installed, the package will be replaced with the current version, Hawk2.
Hawk2's user-friendly Web interface allows you to monitor and administer your High Availability clusters from Linux or non-Linux machines alike. Hawk2 can be accessed from any machine inside or outside of the cluster by using a (graphical) Web browser.
To configure and manage cluster resources, either use the crm shell (crmsh) command line utility or Hawk2, a Web-based user interface.
This chapter introduces crm
, the command line tool
and covers an overview of this tool, how to use templates, and mainly
configuring and managing cluster resources: creating basic and advanced
types of resources (groups and clones), configuring constraints,
specifying failover nodes and failback nodes, configuring resource
monitoring, starting, cleaning up or removing resources, and migrating
resources manually.
All tasks that need to be managed by a cluster must be available as a resource. There are two major groups here to consider: resource agents and STONITH agents. For both categories, you can add your own agents, extending the abilities of the cluster to your own needs.
Fencing is a very important concept in computer clusters for HA (High Availability). A cluster sometimes detects that one of the nodes is behaving strangely and needs to remove it. This is called fencing and is commonly done with a STONITH resource. Fencing may be defined as a method to bring an HA cluster to a known state.
Every resource in a cluster has a state attached. For example: “resource r1 is started on alice”. In an HA cluster, such a state implies that “resource r1 is stopped on all nodes except alice”, because the cluster must make sure that every resource may be started on only one node. Every node must report every change that happens to a resource. The cluster state is thus a collection of resource states and node states.
When the state of a node or resource cannot be established with certainty, fencing comes in. Even when the cluster is not aware of what is happening on a given node, fencing can ensure that the node does not run any important resources.
SBD (STONITH Block Device) provides a node fencing mechanism for Pacemaker-based clusters through the exchange of messages via shared block storage (SAN, iSCSI, FCoE, etc.). This isolates the fencing mechanism from changes in firmware version or dependencies on specific firmware controllers. SBD needs a watchdog on each node to ensure that misbehaving nodes are really stopped. Under certain conditions, it is also possible to use SBD without shared storage, by running it in diskless mode.
The ha-cluster-bootstrap scripts provide an automated way to set up a cluster with the option of using SBD as fencing mechanism. For details, see the Article “Installation and Setup Quick Start”. However, manually setting up SBD provides you with more options regarding the individual settings.
This chapter explains the concepts behind SBD. It guides you through configuring the components needed by SBD to protect your cluster from potential data corruption in case of a split brain scenario.
In addition to node level fencing, you can use additional mechanisms for storage protection, such as LVM2 exclusive activation or OCFS2 file locking support (resource level fencing). They protect your system against administrative or application faults.
The cluster administration tools like crm shell (crmsh) or
Hawk2 can be used by root
or any user in the group
haclient
. By default, these
users have full read/write access. To limit access or assign more
fine-grained access rights, you can use Access control
lists (ACLs).
Access control lists consist of an ordered set of access rules. Each rule allows read or write access or denies access to a part of the cluster configuration. Rules are typically combined to produce a specific role, then users may be assigned to a role that matches their tasks.
For many systems, it is desirable to implement network connections that comply to more than the standard data security or availability requirements of a typical Ethernet device. In these cases, several Ethernet devices can be aggregated to a single bonding device.
Load Balancing makes a cluster of servers appear as one large, fast server to outside clients. This apparent single server is called a virtual server. It consists of one or more load balancers dispatching incoming requests and several real servers running the actual services. With a load balancing setup of High Availability Extension, you can build highly scalable and highly available network services, such as Web, cache, mail, FTP, media and VoIP services.
Apart from local clusters and metro area clusters, SUSE® Linux Enterprise High Availability Extension
15 SP1 also supports geographically dispersed clusters (Geo
clusters, sometimes also called multi-site clusters). That means you can
have multiple, geographically dispersed sites with a local cluster each.
Failover between these clusters is coordinated by a higher level entity,
the so-called booth
. For details on how to
use and set up Geo clusters, refer to Article “Geo Clustering Quick Start” and
Book “Geo Clustering Guide”.
To perform maintenance tasks on the cluster nodes, you might need to stop the resources running on that node, to move them, or to shut down or reboot the node. It might also be necessary to temporarily take over the control of resources from the cluster, or even to stop the cluster service while resources remain running.
This chapter explains how to manually take down a cluster node without negative side-effects. It also gives an overview of different options the cluster stack provides for executing maintenance tasks.
The main purpose of an HA cluster is to manage user services. Typical examples of user services are an Apache Web server or a database. From the user's point of view, the services do something specific when ordered to do so. To the cluster, however, they are only resources which may be started or stopped—the nature of the service is irrelevant to the cluster.
In this chapter, we will introduce some basic concepts you need to know when configuring resources and administering your cluster. The following chapters show you how to execute the main configuration and administration tasks with each of the management tools the High Availability Extension provides.
In general, clusters fall into one of two categories:
Two-node clusters
Clusters with more than two nodes. This usually means an odd number of nodes.
Adding also different topologies, different use cases can be derived. The following use cases are the most common:
Configuration: FC SAN or similar shared storage, layer 2 network.
Usage scenario: Embedded clusters that focus on service high availability and not data redundancy for data replication. Such a setup is used for radio stations or assembly line controllers, for example.
Configuration: Symmetrical stretched cluster, FC SAN, and layer 2 network all across two locations.
Usage scenario: Classic stretched clusters, focus on high availability of services and local data redundancy. For databases and enterprise resource planning. One of the most popular setups during the last few years.
Configuration: 2×N+1 nodes, FC SAN across two main locations. Auxiliary third site with no FC SAN, but acts as a majority maker. Layer 2 network at least across two main locations.
Usage scenario: Classic stretched cluster, focus on high availability of services and data redundancy. For example, databases, enterprise resource planning.
Whenever communication fails between one or more nodes and the rest of the cluster, a cluster partition occurs. The nodes can only communicate with other nodes in the same partition and are unaware of the separated nodes. A cluster partition is defined as having quorum (being “quorate”) if it has the majority of nodes (or votes). How this is achieved is done by quorum calculation. Quorum is a requirement for fencing.
Quorum calculation has changed between SUSE Linux Enterprise High Availability Extension 11 and SUSE Linux Enterprise High Availability Extension 15. For SUSE Linux Enterprise High Availability Extension 11, quorum was calculated by Pacemaker. Starting with SUSE Linux Enterprise High Availability Extension 12, Corosync can handle quorum for two-node clusters directly without changing the Pacemaker configuration.
How quorum is calculated is influenced by the following factors:
To keep services running, a cluster with more than two nodes relies on quorum (majority vote) to resolve cluster partitions. Based on the following formula, you can calculate the minimum number of operational nodes required for the cluster to function:
N ≥ C/2 + 1 N = minimum number of operational nodes C = number of cluster nodes
For example, a five-node cluster needs a minimum of three operational nodes (or two nodes which can fail).
We strongly recommend to use either a two-node cluster or an odd number of cluster nodes. Two-node clusters make sense for stretched setups across two sites. Clusters with an odd number of nodes can either be built on one single site or might be spread across three sites.
Corosync is a messaging and membership layer, see Section 6.2.4, “Corosync Configuration for Two-Node Clusters” and Section 6.2.5, “Corosync Configuration for N-Node Clusters”.
Global cluster options control how the cluster behaves when
confronted with certain situations. They are grouped into sets and can be
viewed and modified with the cluster management tools like Hawk2 and
the crm
shell.
The predefined values can usually be kept. However, to make key functions of your cluster work correctly, you need to adjust the following parameters after basic cluster setup:
no-quorum-policy
#Edit sourceThis global option defines what to do when a cluster partition does not have quorum (no majority of nodes is part of the partition).
Allowed values are:
ignore
Setting no-quorum-policy
to ignore
makes
the cluster behave like it has quorum. Resource management is
continued.
On SLES 11 this was the recommended setting for a two-node cluster. Starting with SLES 12, this option is obsolete. Based on configuration and conditions, Corosync gives cluster nodes or a single node “quorum”—or not.
For two-node clusters the only meaningful behavior is to always react in case of quorum loss. The first step should always be to try to fence the lost node.
freeze
If quorum is lost, the cluster partition freezes. Resource management is continued: running resources are not stopped (but possibly restarted in response to monitor events), but no further resources are started within the affected partition.
This setting is recommended for clusters where certain resources
depend on communication with other nodes (for example, OCFS2 mounts).
In this case, the default setting
no-quorum-policy=stop
is not useful, as it would
lead to the following scenario: Stopping those resources would not be
possible while the peer nodes are unreachable. Instead, an attempt to
stop them would eventually time out and cause a stop
failure
, triggering escalated recovery and fencing.
stop
(default value)If quorum is lost, all resources in the affected cluster partition are stopped in an orderly fashion.
suicide
If quorum is lost, all nodes in the affected cluster partition are fenced. This option works only in combination with SBD, see Chapter 11, Storage Protection and SBD.
stonith-enabled
#Edit source
This global option defines whether to apply fencing, allowing STONITH
devices to shoot failed nodes and nodes with resources that cannot be
stopped. By default, this global option is set to
true
, because for normal cluster operation it is
necessary to use STONITH devices. According to the default value,
the cluster will refuse to start any resources if no STONITH
resources have been defined.
If you need to disable fencing for any reasons, set
stonith-enabled
to false
, but be
aware that this has impact on the support status for your product.
Furthermore, with stonith-enabled="false"
, resources
like the Distributed Lock Manager (DLM) and all services depending on
DLM (such as cLVM, GFS2, and OCFS2) will fail to start.
A cluster without STONITH is not supported.
When using the bootstrap scripts, the Corosync configuration contains
a quorum
section with the following options:
quorum { # Enable and configure quorum subsystem (default: off) # see also corosync.conf.5 and votequorum.5 provider: corosync_votequorum expected_votes: 2 two_node: 1 }
As opposed to SUSE Linux Enterprise 11, the votequorum subsystem in SUSE Linux Enterprise 12 is
powered by Corosync version 2.x. This means that the
no-quorum-policy=ignore
option must not be used.
By default, when two_node: 1
is set, the
wait_for_all
option is automatically enabled.
If wait_for_all
is not enbaled, the cluster should be
started on both nodes in parallel. Otherwise the first node will perform
a startup-fencing on the missing second node.
When not using a two-node cluster, we strongly recommend an odd number of nodes for your N-node cluster. With regards to quorum configuration, you have the following options:
Adding additional nodes with the ha-cluster-join
command, or
Adapting the Corosync configuration manually.
If you adjust /etc/corosync/corosync.conf
manually,
use the following settings:
quorum { provider: corosync_votequorum 1 expected_votes: N 2 wait_for_all: 1 3 }
Use the quorum service from Corosync | |
The number of votes to expect. This parameter can either be
provided inside the | |
Enables the wait for all (WFA) feature.
When WFA is enabled, the cluster will be quorate for the first time
only after all nodes have become visible.
To avoid some startup race conditions, setting |
As a cluster administrator, you need to create cluster resources for every resource or application you run on servers in your cluster. Cluster resources can include Web sites, e-mail servers, databases, file systems, virtual machines, and any other server-based applications or services you want to make available to users at all times.
Before you can use a resource in the cluster, it must be set up. For example, to use an Apache server as a cluster resource, set up the Apache server first and complete the Apache configuration before starting the respective resource in your cluster.
If a resource has specific environment requirements, make sure they are present and identical on all cluster nodes. This kind of configuration is not managed by the High Availability Extension. You must do this yourself.
When managing a resource with the High Availability Extension, the same resource must not be started or stopped otherwise (outside of the cluster, for example manually or on boot or reboot). The High Availability Extension software is responsible for all service start or stop actions.
If you need to execute testing or maintenance tasks after the services are already running under cluster control, make sure to put the resources, nodes, or the whole cluster into maintenance mode before you touch any of them manually. For details, see Section 16.2, “Different Options for Maintenance Tasks”.
After having configured the resources in the cluster, use the cluster management tools to start, stop, clean up, remove or migrate any resources manually. For details how to do so with your preferred cluster management tool:
For each cluster resource you add, you need to define the standard that the resource agent conforms to. Resource agents abstract the services they provide and present an accurate status to the cluster, which allows the cluster to be non-committal about the resources it manages. The cluster relies on the resource agent to react appropriately when given a start, stop or monitor command.
Typically, resource agents come in the form of shell scripts. The High Availability Extension supports the following classes of resource agents:
OCF RA agents are best suited for use with High Availability, especially when
you need promotable clone resources or special monitoring abilities. The
agents are generally located in
/usr/lib/ocf/resource.d/provider/
.
Their functionality is similar to that of LSB scripts. However, the
configuration is always done with environmental variables which allow
them to accept and process parameters easily.
OCF specifications have strict definitions of which exit codes must
be returned by actions, see Section 9.3, “OCF Return Codes and Failure Recovery”. The
cluster follows these specifications exactly.
All OCF Resource Agents are required to have at least the actions
start
, stop
,
status
, monitor
, and
meta-data
. The meta-data
action
retrieves information about how to configure the agent. For example,
to know more about the IPaddr
agent by
the provider heartbeat
, use the following command:
OCF_ROOT=/usr/lib/ocf /usr/lib/ocf/resource.d/heartbeat/IPaddr meta-data
The output is information in XML format, including several sections (general description, available parameters, available actions for the agent).
Alternatively, use the crmsh to view information on OCF resource agents. For details, see Section 8.1.3, “Displaying Information about OCF Resource Agents”.
LSB resource agents are generally provided by the operating
system/distribution and are found in
/etc/init.d
. To be used with the cluster, they
must conform to the LSB init script specification. For example, they
must have several actions implemented, which are, at minimum,
start
, stop
,
restart
, reload
,
force-reload
, and status
. For
more information, see
http://refspecs.linuxbase.org/LSB_4.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html.
The configuration of those services is not standardized. If you
intend to use an LSB script with High Availability, make sure that you
understand how the relevant script is configured. Often you can find
information about this in the documentation of the relevant package
in
/usr/share/doc/packages/PACKAGENAME
.
Starting with SUSE Linux Enterprise 12, systemd is a replacement for the popular System V init daemon. Pacemaker can manage systemd services if they are present. Instead of init scripts, systemd has unit files. Generally the services (or unit files) are provided by the operating system. In case you want to convert existing init scripts, find more information at http://0pointer.de/blog/projects/systemd-for-admins-3.html.
There are currently many “common” types of system
services that exist in parallel: LSB
(belonging to
System V init), systemd
, and (in some
distributions) upstart
. Therefore, Pacemaker
supports a special alias which intelligently figures out which one
applies to a given cluster node. This is particularly useful when the
cluster contains a mix of systemd, upstart, and LSB services.
Pacemaker will try to find the named service in the following order:
as an LSB (SYS-V) init script, a systemd unit file, or an Upstart
job.
Monitoring plug-ins (formerly called Nagios plug-ins) allow to monitor services on remote hosts. Pacemaker can do remote monitoring with the monitoring plug-ins if they are present. For detailed information, see Section 6.6.1, “Monitoring Services on Remote Hosts with Monitoring Plug-ins”.
This class is used exclusively for fencing related resources. For more information, see Chapter 10, Fencing and STONITH.
The agents supplied with the High Availability Extension are written to OCF specifications.
The following types of resources can be created:
A primitive resource, the most basic type of resource.
Learn how to create primitive resources with your preferred cluster management tool:
Groups contain a set of resources that need to be located together, started sequentially and stopped in the reverse order. For more information, refer to Section 6.3.5.1, “Groups”.
Clones are resources that can be active on multiple hosts. Any resource can be cloned, provided the respective resource agent supports it. For more information, refer to Section 6.3.5.2, “Clones”.
Promotable clones (formerly known master/slave or multi-state resources) are a special type of clone resources that can be promoted.
If you want to create lots of resources with similar configurations, defining a resource template is the easiest way. After having been defined, it can be referenced in primitives—or in certain types of constraints, as described in Section 6.5.3, “Resource Templates and Constraints”.
If a template is referenced in a primitive, the primitive will inherit all operations, instance attributes (parameters), meta attributes, and utilization attributes defined in the template. Additionally, you can define specific operations or attributes for your primitive. If any of these are defined in both the template and the primitive, the values defined in the primitive will take precedence over the ones defined in the template.
Learn how to define resource templates with your preferred cluster configuration tool:
Whereas primitives are the simplest kind of resources and therefore easy to configure, you will probably also need more advanced resource types for cluster configuration, such as groups, clones or promotable clone resources.
Some cluster resources depend on other components or resources. They require that each component or resource starts in a specific order and runs together on the same server with resources it depends on. To simplify this configuration, you can use cluster resource groups.
An example of a resource group would be a Web server that requires an IP address and a file system. In this case, each component is a separate resource that is combined into a cluster resource group. The resource group would run on one or more servers. In case of a software or hardware malfunction, the group would fail over to another server in the cluster, similar to an individual cluster resource.
Groups have the following properties:
Resources are started in the order they appear in and stopped in the reverse order.
If a resource in the group cannot run anywhere, then none of the resources located after that resource in the group is allowed to run.
Groups may only contain a collection of primitive cluster resources. Groups must contain at least one resource, otherwise the configuration is not valid. To refer to the child of a group resource, use the child’s ID instead of the group’s ID.
Although it is possible to reference the group’s children in constraints, it is usually preferable to use the group’s name instead.
Stickiness is additive in groups. Every active
member of the group will contribute its stickiness value to the
group’s total. So if the default
resource-stickiness
is 100
and
a group has seven members (five of which are active), the group as
a whole will prefer its current location with a score of
500
.
To enable resource monitoring for a group, you must configure monitoring separately for each resource in the group that you want monitored.
Learn how to create groups with your preferred cluster management tool:
You may want certain resources to run simultaneously on multiple nodes in your cluster. To do this you must configure a resource as a clone. Examples of resources that might be configured as clones include cluster file systems like OCFS2. You can clone any resource provided. This is supported by the resource’s Resource Agent. Clone resources may even be configured differently depending on which nodes they are hosted.
There are three types of resource clones:
These are the simplest type of clones. They behave identically anywhere they are running. Because of this, there can only be one instance of an anonymous clone active per machine.
These resources are distinct entities. An instance of the clone running on one node is not equivalent to another instance on another node; nor would any two instances on the same node be equivalent.
Active instances of these resources are divided into two states, active and passive. These are also sometimes called primary and secondary, or master and slave. Promotable clones can be either anonymous or globally unique. See also Section 6.3.5.3, “Promotable Clones (Multi-state Resources)”.
Clones must contain exactly one group or one regular resource.
When configuring resource monitoring or constraints, clones have different requirements than simple resources. For details, see Pacemaker Explained, available from http://www.clusterlabs.org/pacemaker/doc/. Refer to section Clones - Resources That Get Active on Multiple Hosts.
Learn how to create clones with your preferred cluster management tool:
Promotable clones (formerly known as multi-state resources) are a
specialization of clones. They allow the
instances to be in one of two operating modes (called
master
or slave
, but can mean
whatever you want them to mean). Promotable clones must contain
exactly one group or one regular resource.
When configuring resource monitoring or constraints, promotable clones have different requirements than simple resources. For details, see Pacemaker Explained. The version for Pacemaker 1.1 is available from http://www.clusterlabs.org/pacemaker/doc/. Refer to section Multi-state - Resources That Have Multiple Modes.
For each resource you add, you can define options. Options are used by
the cluster to decide how your resource should behave—they tell
the CRM how to treat a specific resource. Resource options can be set
with the crm_resource --meta
command or with
Hawk2 as described in
Procedure 7.5, “Adding a Primitive Resource”.
Option |
Description |
Default |
---|---|---|
|
If not all resources can be active, the cluster will stop lower priority resources to keep higher priority ones active. |
|
|
In what state should the cluster attempt to keep this resource?
Allowed values: |
|
|
Is the cluster allowed to start and stop the resource? Allowed
values: |
|
|
Can the resources be touched manually? Allowed values:
|
|
|
How much does the resource prefer to stay where it is? |
calculated |
|
How many failures should occur for this resource on a node before making the node ineligible to host this resource? |
|
|
What should the cluster do if it ever finds the resource active on
more than one node? Allowed values: |
|
|
How many seconds to wait before acting as if the failure had not occurred (and potentially allowing the resource back to the node on which it failed)? |
|
|
Allow resource migration for resources which support
|
|
|
The name of the remote node this resource defines. This both
enables the resource as a remote node and defines the unique name
used to identify the remote node. If no other parameters are set,
this value will also be assumed as the host name to connect to at
![]() Warning: Use Unique IDsThis value must not overlap with any existing resource or node IDs. |
none (disabled) |
|
Custom port for the guest connection to pacemaker_remote. |
|
|
The IP address or host name to connect to if the remote node’s name is not the host name of the guest. |
|
|
How long before a pending guest connection will time out. |
|
The scripts of all resource classes can be given parameters which
determine how they behave and which instance of a service they control.
If your resource agent supports parameters, you can add them with the
crm_resource
command or with
Hawk2 as described in
Procedure 7.5, “Adding a Primitive Resource”. In the
crm
command line utility and in Hawk2, instance
attributes are called params
or
Parameter
, respectively. The list of instance
attributes supported by an OCF script can be found by executing the
following command as root
:
root #
crm
ra info [class:[provider:]]resource_agent
or (without the optional parts):
root #
crm
ra info resource_agent
The output lists all the supported attributes, their purpose and default values.
For example, the command
root #
crm
ra info IPaddr
returns the following output:
Manages virtual IPv4 addresses (portable version) (ocf:heartbeat:IPaddr) This script manages IP alias IP addresses It can add an IP alias, or remove one. Parameters (* denotes required, [] the default): ip* (string): IPv4 address The IPv4 address to be configured in dotted quad notation, for example "192.168.1.1". nic (string, [eth0]): Network interface The base network interface on which the IP address will be brought online. If left empty, the script will try and determine this from the routing table. Do NOT specify an alias interface in the form eth0:1 or anything here; rather, specify the base interface only. cidr_netmask (string): Netmask The netmask for the interface in CIDR format. (ie, 24), or in dotted quad notation 255.255.255.0). If unspecified, the script will also try to determine this from the routing table. broadcast (string): Broadcast address Broadcast address associated with the IP. If left empty, the script will determine this from the netmask. iflabel (string): Interface label You can specify an additional label for your IP address here. lvs_support (boolean, [false]): Enable support for LVS DR Enable support for LVS Direct Routing configurations. In case a IP address is stopped, only move it to the loopback device to allow the local node to continue to service requests, but no longer advertise it on the network. local_stop_script (string): Script called when the IP is released local_start_script (string): Script called when the IP is added ARP_INTERVAL_MS (integer, [500]): milliseconds between gratuitous ARPs milliseconds between ARPs ARP_REPEAT (integer, [10]): repeat count How many gratuitous ARPs to send out when bringing up a new address ARP_BACKGROUND (boolean, [yes]): run in background run in background (no longer any reason to do this) ARP_NETMASK (string, [ffffffffffff]): netmask for ARP netmask for ARP - in nonstandard hexadecimal format. Operations' defaults (advisory minimum): start timeout=90 stop timeout=100 monitor_0 interval=5s timeout=20s
Note that groups, clones and promotable clone resources do not have instance attributes. However, any instance attributes set will be inherited by the group's, clone's or promotable clone's children.
By default, the cluster will not ensure that your resources are still healthy. To instruct the cluster to do this, you need to add a monitor operation to the resource’s definition. Monitor operations can be added for all classes or resource agents. For more information, refer to Section 6.4, “Resource Monitoring”.
Operation |
Description |
---|---|
|
Your name for the action. Must be unique. (The ID is not shown). |
|
The action to perform. Common values: |
|
How frequently to perform the operation. Unit: seconds |
|
How long to wait before declaring the action has failed. |
|
What conditions need to be satisfied before this action occurs.
Allowed values: |
|
The action to take if this action ever fails. Allowed values:
|
|
If |
|
Run the operation only if the resource has this role. |
|
Can be set either globally or for individual resources. Makes the CIB reflect the state of “in-flight” operations on resources. |
|
Description of the operation. |
Timeouts values for resources can be influenced by the following parameters:
op_defaults
(global timeout for operations),
a specific timeout value defined in a resource template,
a specific timeout value defined for a resource.
If a specific value is defined for a resource, it takes precedence over the global default. A specific value for a resource also takes precedence over a value that is defined in a resource template.
Getting timeout values right is very important. Setting them too low will result in a lot of (unnecessary) fencing operations for the following reasons:
If a resource runs into a timeout, it fails and the cluster will try to stop it.
If stopping the resource also fails (for example, because the timeout for stopping is set too low), the cluster will fence the node. It considers the node where this happens to be out of control.
You can adjust the global default for operations and set any specific timeout values with both crmsh and Hawk2. The best practice for determining and setting timeout values is as follows:
Check how long it takes your resources to start and stop (under load).
If needed, add the op_defaults
parameter and set
the (default) timeout value accordingly:
For example, set op_defaults
to
60
seconds:
crm(live)configure#
op_defaults timeout=60
For resources that need longer periods of time, define individual timeout values.
When configuring operations for a resource, add separate
start
and stop
operations. When
configuring operations with Hawk2, it will provide useful timeout
proposals for those operations.
If you want to ensure that a resource is running, you must configure resource monitoring for it.
If the resource monitor detects a failure, the following takes place:
Log file messages are generated, according to the configuration
specified in the logging
section of
/etc/corosync/corosync.conf
.
The failure is reflected in the cluster management tools (Hawk2,
crm status
), and in the CIB status section.
The cluster initiates noticeable recovery actions which may include stopping the resource to repair the failed state and restarting the resource locally or on another node. The resource also may not be restarted, depending on the configuration and state of the cluster.
If you do not configure resource monitoring, resource failures after a successful start will not be communicated, and the cluster will always show the resource as healthy.
Usually, resources are only monitored by the cluster as long as they are running. However, to detect concurrency violations, also configure monitoring for resources which are stopped. For example:
primitive dummy1 ocf:heartbeat:Dummy \ op monitor interval="300s" role="Stopped" timeout="10s" \ op monitor interval="30s" timeout="10s"
This configuration triggers a monitoring operation every
300
seconds for the resource
dummy1
when it is in
role="Stopped"
. When running, it will be monitored
every 30
seconds.
The CRM executes an initial monitoring for each resource on every
node, the so-called probe
. A probe is also executed
after the cleanup of a resource. If multiple monitoring operations are
defined for a resource, the CRM will select the one with the smallest
interval and will use its timeout value as default timeout for
probing. If no monitor operation is configured, the cluster-wide
default applies. The default is 20
seconds (if not
specified otherwise by configuring the op_defaults
parameter). If you do not want to rely on the automatic calculation or
the op_defaults
value, define a specific
monitoring operation for the probing of this
resource. Do so by adding a monitoring operation with the
interval
set to 0
, for example:
crm(live)configure#
primitive
rsc1 ocf:pacemaker:Dummy \ op monitor interval="0" timeout="60"
The probe of rsc1
will time out in
60s
, independent of the global timeout defined in
op_defaults
, or any other operation timeouts
configured. If you did not set interval="0"
for
specifying the probing of the respective resource, the CRM will
automatically check for any other monitoring operations defined for
that resource and will calculate the timeout value for probing as
described above.
Learn how to add monitor operations to resources with your preferred cluster management tool:
Having all the resources configured is only part of the job. Even if the cluster knows all needed resources, it might still not be able to handle them correctly. Resource constraints let you specify which cluster nodes resources can run on, what order resources will load, and what other resources a specific resource is dependent on.
There are three different kinds of constraints available:
Locational constraints that define on which nodes a resource may be run, may not be run or is preferred to be run.
Colocational constraints that tell the cluster which resources may or may not run together on a node.
Ordering constraints to define the sequence of actions.
Do not create colocation constraints for members of a resource group. Create a colocation constraint pointing to the resource group as a whole instead. All other types of constraints are safe to use for members of a resource group.
Do not use any constraints on a resource that has a clone resource or a promotable clone resource applied to it. The constraints must apply to the clone or promotable clone resource, not to the child resource.
As an alternative format for defining location, colocation or ordering
constraints, you can use resource sets
, where
primitives are grouped together in one set. Previously this was
possible either by defining a resource group (which could not always
accurately express the design), or by defining each relationship as an
individual constraint. The latter caused a constraint explosion as the
number of resources and combinations grew. The configuration via
resource sets is not necessarily less verbose, but is easier to
understand and maintain, as the following examples show.
For example, you can use the following configuration of a resource
set (loc-alice
) in the crmsh to place
two virtual IPs (vip1
and vip2
)
on the same node, alice
:
crm(live)configure#
primitive
vip1 ocf:heartbeat:IPaddr2 params ip=192.168.1.5crm(live)configure#
primitive
vip1 ocf:heartbeat:IPaddr2 params ip=192.168.1.6crm(live)configure#
location
loc-alice { vip1 vip2 } inf: alice
If you want to use resource sets to replace a configuration of colocation constraints, consider the following two examples:
<constraints> <rsc_colocation id="coloc-1" rsc="B" with-rsc="A" score="INFINITY"/> <rsc_colocation id="coloc-2" rsc="C" with-rsc="B" score="INFINITY"/> <rsc_colocation id="coloc-3" rsc="D" with-rsc="C" score="INFINITY"/> </constraints>
The same configuration expressed by a resource set:
<constraints> <rsc_colocation id="coloc-1" score="INFINITY" > <resource_set id="colocated-set-example" sequential="true"> <resource_ref id="A"/> <resource_ref id="B"/> <resource_ref id="C"/> <resource_ref id="D"/> </resource_set> </rsc_colocation> </constraints>
If you want to use resource sets to replace a configuration of ordering constraints, consider the following two examples:
<constraints> <rsc_order id="order-1" first="A" then="B" /> <rsc_order id="order-2" first="B" then="C" /> <rsc_order id="order-3" first="C" then="D" /> </constraints>
The same purpose can be achieved by using a resource set with ordered resources:
<constraints> <rsc_order id="order-1"> <resource_set id="ordered-set-example" sequential="true"> <resource_ref id="A"/> <resource_ref id="B"/> <resource_ref id="C"/> <resource_ref id="D"/> </resource_set> </rsc_order> </constraints>
Sets can be either ordered (sequential=true
) or
unordered (sequential=false
). Furthermore, the
require-all
attribute can be used to switch between
AND
and OR
logic.
Sometimes it is useful to place a group of resources on the same node
(defining a colocation constraint), but without having hard
dependencies between the resources. For example, you want two
resources to be placed on the same node, but you do
not want the cluster to restart the other one if
one of them fails. This can be achieved on the crm shell by using
the weak bond
command.
Learn how to set these “weak bonds” with your preferred cluster management tool:
Learn how to add the various kinds of constraints with your preferred cluster management tool:
For more information on configuring constraints and detailed background information about the basic concepts of ordering and colocation, refer to the following documents. They are available at http://www.clusterlabs.org/pacemaker/doc/:
Pacemaker Explained, chapter Resource Constraints
Colocation Explained
Ordering Explained
When defining constraints, you also need to deal with scores. Scores of all kinds are integral to how the cluster works. Practically everything from migrating a resource to deciding which resource to stop in a degraded cluster is achieved by manipulating scores in some way. Scores are calculated on a per-resource basis and any node with a negative score for a resource cannot run that resource. After calculating the scores for a resource, the cluster then chooses the node with the highest score.
INFINITY
is currently defined as
1,000,000
. Additions or subtractions with it stick to
the following three basic rules:
Any value + INFINITY = INFINITY
Any value - INFINITY = -INFINITY
INFINITY - INFINITY = -INFINITY
When defining resource constraints, you specify a score for each constraint. The score indicates the value you are assigning to this resource constraint. Constraints with higher scores are applied before those with lower scores. By creating additional location constraints with different scores for a given resource, you can specify an order for the nodes that a resource will fail over to.
If you have defined a resource template (see Section 6.3.4, “Resource Templates”), it can be referenced in the following types of constraints:
order constraints,
colocation constraints,
rsc_ticket constraints (for Geo clusters).
However, colocation constraints must not contain more than one reference to a template. Resource sets must not contain a reference to a template.
Resource templates referenced in constraints stand for all primitives which are derived from that template. This means, the constraint applies to all primitive resources referencing the resource template. Referencing resource templates in constraints is an alternative to resource sets and can simplify the cluster configuration considerably. For details about resource sets, refer to Procedure 7.17, “Using a Resource Set for Constraints”.
A resource will be automatically restarted if it fails. If that cannot
be achieved on the current node, or it fails N
times
on the current node, it will try to fail over to another node. Each time
the resource fails, its failcount is raised. You can define several
failures for resources (a migration-threshold
), after
which they will migrate to a new node. If you have more than two nodes
in your cluster, the node a particular resource fails over to is chosen
by the High Availability software.
However, you can specify the node a resource will fail over to by
configuring one or several location constraints and a
migration-threshold
for that resource.
Learn how to specify failover nodes with your preferred cluster management tool:
For example, let us assume you have configured a location constraint
for resource rsc1
to preferably run on
alice
. If it fails there,
migration-threshold
is checked and compared to the
failcount. If failcount >= migration-threshold then the resource is
migrated to the node with the next best preference.
After the threshold has been reached, the node will no longer be
allowed to run the failed resource until the resource's failcount is
reset. This can be done manually by the cluster administrator or by
setting a failure-timeout
option for the resource.
For example, a setting of migration-threshold=2
and
failure-timeout=60s
would cause the resource to
migrate to a new node after two failures. It would be allowed to move
back (depending on the stickiness and constraint scores) after one
minute.
There are two exceptions to the migration threshold concept, occurring when a resource either fails to start or fails to stop:
Start failures set the failcount to INFINITY
and
thus always cause an immediate migration.
Stop failures cause fencing (when stonith-enabled
is set to true
which is the default).
In case there is no STONITH resource defined (or
stonith-enabled
is set to
false
), the resource will not migrate.
For details on using migration thresholds and resetting failcounts with your preferred cluster management tool:
A resource might fail back to its original node when that node is back online and in the cluster. To prevent a resource from failing back to the node that it was running on, or to specify a different node for the resource to fail back to, change its resource stickiness value. You can either specify resource stickiness when you are creating a resource or afterward.
Consider the following implications when specifying resource stickiness values:
0
:This is the default. The resource will be placed optimally in the system. This may mean that it is moved when a “better” or less loaded node becomes available. This option is almost equivalent to automatic failback, except that the resource may be moved to a node that is not the one it was previously active on.
0
:The resource will prefer to remain in its current location, but may be moved if a more suitable node is available. Higher values indicate a stronger preference for a resource to stay where it is.
0
:The resource prefers to move away from its current location. Higher absolute values indicate a stronger preference for a resource to be moved.
INFINITY
:
The resource will always remain in its current location unless forced
off because the node is no longer eligible to run the resource (node
shutdown, node standby, reaching the
migration-threshold
, or configuration change).
This option is almost equivalent to completely disabling automatic
failback.
-INFINITY
:The resource will always move away from its current location.
Not all resources are equal. Some, such as Xen guests, require that the node hosting them meets their capacity requirements. If resources are placed such that their combined need exceed the provided capacity, the resources diminish in performance (or even fail).
To take this into account, the High Availability Extension allows you to specify the following parameters:
The capacity a certain node provides.
The capacity a certain resource requires.
An overall strategy for placement of resources.
Learn how to configure these settings with your preferred cluster management tool:
A node is considered eligible for a resource if it has sufficient free capacity to satisfy the resource's requirements. The nature of the capacities is completely irrelevant for the High Availability Extension; it only makes sure that all capacity requirements of a resource are satisfied before moving a resource to a node.
To manually configure the resource's requirements and the capacity a node provides, use utilization attributes. You can name the utilization attributes according to your preferences and define as many name/value pairs as your configuration needs. However, the attribute's values must be integers.
If multiple resources with utilization attributes are grouped or have colocation constraints, the High Availability Extension takes that into account. If possible, the resources will be placed on a node that can fulfill all capacity requirements.
It is impossible to set utilization attributes directly for a resource group. However, to simplify the configuration for a group, you can add a utilization attribute with the total capacity needed to any of the resources in the group.
The High Availability Extension also provides means to detect and configure both node capacity and resource requirements automatically:
The NodeUtilization
resource agent checks the
capacity of a node (regarding CPU and RAM).
To configure automatic detection, create a clone resource of the
following class, provider, and type:
ocf:pacemaker:NodeUtilization
. One instance of the
clone should be running on each node. After the instance has started, a
utilization section will be added to the node's configuration in CIB.
For automatic detection of a resource's minimal requirements (regarding
RAM and CPU) the Xen
resource agent has been
improved. Upon start of a Xen
resource, it will
reflect the consumption of RAM and CPU. Utilization attributes will
automatically be added to the resource configuration.
The ocf:heartbeat:Xen
resource agent should not be
used with libvirt
, as libvirt
expects
to be able to modify the machine description file.
For libvirt
, use the
ocf:heartbeat:VirtualDomain
resource agent.
Apart from detecting the minimal requirements, the High Availability Extension also allows
to monitor the current utilization via the
VirtualDomain
resource agent. It detects CPU
and RAM use of the virtual machine. To use this feature, configure a
resource of the following class, provider and type:
ocf:heartbeat:VirtualDomain
. The following instance
attributes are available: autoset_utilization_cpu
and
autoset_utilization_hv_memory
. Both default to
true
. This updates the utilization values in the CIB
during each monitoring cycle.
Independent of manually or automatically configuring capacity and
requirements, the placement strategy must be specified with the
placement-strategy
property (in the global cluster
options). The following values are available:
default
(default value)Utilization values are not considered. Resources are allocated according to location scoring. If scores are equal, resources are evenly distributed across nodes.
utilization
Utilization values are considered when deciding if a node has enough free capacity to satisfy a resource's requirements. However, load-balancing is still done based on the number of resources allocated to a node.
minimal
Utilization values are considered when deciding if a node has enough free capacity to satisfy a resource's requirements. An attempt is made to concentrate the resources on as few nodes as possible (to achieve power savings on the remaining nodes).
balanced
Utilization values are considered when deciding if a node has enough free capacity to satisfy a resource's requirements. An attempt is made to distribute the resources evenly, thus optimizing resource performance.
The available placement strategies are best-effort—they do not yet use complex heuristic solvers to always reach optimum allocation results. Ensure that resource priorities are properly set so that your most important resources are scheduled first.
The following example demonstrates a three-node cluster of equal nodes, with four virtual machines.
node alice utilization memory="4000" node bob utilization memory="4000" node charlie utilization memory="4000" primitive xenA ocf:heartbeat:Xen utilization hv_memory="3500" \ params xmfile="/etc/xen/shared-vm/vm1" meta priority="10" primitive xenB ocf:heartbeat:Xen utilization hv_memory="2000" \ params xmfile="/etc/xen/shared-vm/vm2" meta priority="1" primitive xenC ocf:heartbeat:Xen utilization hv_memory="2000" \ params xmfile="/etc/xen/shared-vm/vm3" meta priority="1" primitive xenD ocf:heartbeat:Xen utilization hv_memory="1000" \ params xmfile="/etc/xen/shared-vm/vm4" meta priority="5" property placement-strategy="minimal"
With all three nodes up, resource xenA
will be
placed onto a node first, followed by xenD
.
xenB
and xenC
would either be
allocated together or one of them with xenD
.
If one node failed, too little total memory would be available to host
them all. xenA
would be ensured to be allocated, as
would xenD
. However, only one of the remaining
resources xenB
or xenC
could
still be placed. Since their priority is equal, the result would still
be open. To resolve this ambiguity as well, you would need to set a
higher priority for either one.
Tags are a new feature that has been added to Pacemaker recently. Tags
are a way to refer to multiple resources at once, without creating any
colocation or ordering relationship between them. This can be useful for
grouping conceptually related resources. For example, if you have
several resources related to a database, create a tag called
databases
and add all resources related to the
database to this tag. This allows you to stop or start them all with a
single command.
Tags can also be used in constraints. For example, the following
location constraint loc-db-prefer
applies to the set
of resources tagged with databases
:
location loc-db-prefer databases 100: alice
Learn how to create tags with your preferred cluster management tool:
The possibilities for monitoring and managing services on remote hosts
has become increasingly important during the last few years.
SUSE Linux Enterprise High Availability Extension 11 SP3 offered fine-grained monitoring of services on
remote hosts via monitoring plug-ins. The recent addition of the
pacemaker_remote
service now allows SUSE Linux Enterprise High Availability Extension
15 SP1 to fully manage and monitor resources on remote hosts
just as if they were a real cluster node—without the need to
install the cluster stack on the remote machines.
Monitoring of virtual machines can be done with the VM agent (which only checks if the guest shows up in the hypervisor), or by external scripts called from the VirtualDomain or Xen agent. Up to now, more fine-grained monitoring was only possible with a full setup of the High Availability stack within the virtual machines.
By providing support for monitoring plug-ins (formerly named Nagios plug-ins), the High Availability Extension now also allows you to monitor services on remote hosts. You can collect external statuses on the guests without modifying the guest image. For example, VM guests might run Web services or simple network resources that need to be accessible. With the Nagios resource agents, you can now monitor the Web service or the network resource on the guest. In case these services are not reachable anymore, the High Availability Extension will trigger a restart or migration of the respective guest.
If your guests depend on a service (for example, an NFS server to be used by the guest), the service can either be an ordinary resource, managed by the cluster, or an external service that is monitored with Nagios resources instead.
To configure the Nagios resources, the following packages must be installed on the host:
monitoring-plugins
monitoring-plugins-metadata
YaST or Zypper will resolve any dependencies on further packages, if required.
A typical use case is to configure the monitoring plug-ins as resources belonging to a resource container, which usually is a VM. The container will be restarted if any of its resources has failed. Refer to Example 6.10, “Configuring Resources for Monitoring Plug-ins” for a configuration example. Alternatively, Nagios resource agents can also be configured as ordinary resources to use them for monitoring hosts or services via the network.
primitive vm1 ocf:heartbeat:VirtualDomain \ params hypervisor="qemu:///system" config="/etc/libvirt/qemu/vm1.xml" \ op start interval="0" timeout="90" \ op stop interval="0" timeout="90" \ op monitor interval="10" timeout="30" primitive vm1-sshd nagios:check_tcp \ params hostname="vm1" port="22" \ 1 op start interval="0" timeout="120" \ 2 op monitor interval="10" group g-vm1-and-services vm1 vm1-sshd \ meta container="vm1" 3
The supported parameters are the same as the long options of a
monitoring plug-in. Monitoring plug-ins connect to services with the
parameter | |
As it takes some time to get the guest operating system up and its services running, the start timeout of the monitoring resource must be long enough. | |
A cluster resource container of type
|
The example above contains only one resource for the
check_tcp
plug-in, but multiple resources for
different plug-in types can be configured (for example,
check_http
or check_udp
).
If the host names of the services are the same, the
hostname
parameter can also be specified for the
group, instead of adding it to the individual primitives. For example:
group g-vm1-and-services vm1 vm1-sshd vm1-httpd \ meta container="vm1" \ params hostname="vm1"
If any of the services monitored by the monitoring plug-ins fail within
the VM, the cluster will detect that and restart the container resource
(the VM). Which action to take in this case can be configured by
specifying the on-fail
attribute for the service's
monitoring operation. It defaults to
restart-container
.
Failure counts of services will be taken into account when considering the VM's migration-threshold.
pacemaker_remote
#Edit source
With the pacemaker_remote
service, High Availability clusters
can be extended to virtual nodes or remote bare-metal machines. They do
not need to run the cluster stack to become members of the cluster.
The High Availability Extension can now launch virtual environments (KVM and LXC), plus the resources that live within those virtual environments without requiring the virtual environments to run Pacemaker or Corosync.
For the use case of managing both virtual machines as cluster resources plus the resources that live within the VMs, you can now use the following setup:
The “normal” (bare-metal) cluster nodes run the High Availability Extension.
The virtual machines run the pacemaker_remote
service (almost no configuration required on the VM's side).
The cluster stack on the “normal” cluster nodes launches
the VMs and connects to the pacemaker_remote
service running on the VMs to integrate them as remote nodes into the
cluster.
As the remote nodes do not have the cluster stack installed, this has the following implications:
Remote nodes do not take part in quorum.
Remote nodes cannot become the DC.
Remote nodes are not bound by the scalability limits (Corosync has a member limit of 32 nodes).
Find more information about the remote_pacemaker
service, including multiple use cases with detailed setup instructions
in Article “Pacemaker Remote Quick Start”.
To prevent a node from running out of disk space and thus being unable to
manage any resources that have been assigned to it, the High Availability Extension
provides a resource agent,
ocf:pacemaker:SysInfo
. Use it to monitor a
node's health with regard to disk partitions.
The SysInfo RA creates a node attribute named
#health_disk
which will be set to
red
if any of the monitored disks' free space is below
a specified limit.
To define how the CRM should react in case a node's health reaches a
critical state, use the global cluster option
node-health-strategy
.
To automatically move resources away from a node in case the node runs out of disk space, proceed as follows:
Configure an ocf:pacemaker:SysInfo
resource:
primitive sysinfo ocf:pacemaker:SysInfo \ params disks="/tmp /var"1 min_disk_free="100M"2 disk_unit="M"3 \ op monitor interval="15s"
Which disk partitions to monitor. For example,
![]() Note: | |
The minimum free disk space required for those partitions.
Optionally, you can specify the unit to use for measurement (in the
example above, | |
The unit in which to report the disk space. |
To complete the resource configuration, create a clone of
ocf:pacemaker:SysInfo
and start it on each
cluster node.
Set the node-health-strategy
to
migrate-on-red
:
property node-health-strategy="migrate-on-red"
In case of a #health_disk
attribute set to
red
, the pacemaker-schedulerd
adds -INF
to the resources' score for that node. This will cause any resources to
move away from this node. The STONITH resource will be the last
one to be stopped but even if the STONITH resource is not running
anymore, the node can still be fenced. Fencing has direct access to the
CIB and will continue to work.
After a node's health status has turned to red
, solve
the issue that led to the problem. Then clear the red
status to make the node eligible again for running resources. Log in to
the cluster node and use one of the following methods:
Execute the following command:
root #
crm
node status-attr NODE delete #health_disk
Restart Pacemaker on that node.
Reboot the node.
The node will be returned to service and can run resources again.
Home page of the crm shell (crmsh), the advanced command line interface for High Availability cluster management.
Holds several documents about the crm shell, including a Getting Started tutorial for basic cluster setup with crmsh and the comprehensive Manual for the crm shell. The latter is available at http://crmsh.github.io/man-2.0/. Find the tutorial at http://crmsh.github.io/start-guide/.
Home page of Pacemaker, the cluster resource manager shipped with the High Availability Extension.
Holds several comprehensive manuals and some shorter documents explaining general concepts. For example:
Pacemaker Explained: Contains comprehensive and very detailed information for reference.
Colocation Explained
Ordering Explained
To configure and manage cluster resources, either use Hawk2, or the crm shell (crmsh) command line utility. If you upgrade from an earlier version of SUSE® Linux Enterprise High Availability Extension where Hawk was installed, the package will be replaced with the current version, Hawk2.
Hawk2's user-friendly Web interface allows you to monitor and administer your High Availability clusters from Linux or non-Linux machines alike. Hawk2 can be accessed from any machine inside or outside of the cluster by using a (graphical) Web browser.
Before users can log in to Hawk2, the following requirements need to be fulfilled:
The hawk2 package must be installed on all cluster nodes you want to connect to with Hawk2.
On the machine from which to access a cluster node using Hawk2, you need a (graphical) Web browser (with JavaScript and cookies enabled) to establish the connection.
To use Hawk2, the respective Web service must be started on the node that you want to connect to via the Web interface. See Procedure 7.1, “Starting Hawk2 Services”.
If you have set up your cluster with the scripts from the
ha-cluster-bootstrap
package,
the Hawk2 service is already enabled.
Hawk2 users must be members of the haclient
group. The installation creates a
Linux user named hacluster
, who
is added to the haclient
group.
When using the ha-cluster-init
script for setup,
a default password is set for the hacluster
user. Before starting Hawk2, change it to a
secure password. If you did not use the ha-cluster-init
script, either set a password for the hacluster
first or create a new user which is a member of
the haclient
group. Do this on
every node you will connect to with Hawk2.
On the node you want to connect to, open a shell and log in as root
.
Check the status of the service by entering
root #
systemctl
status hawk
If the service is not running, start it with
root #
systemctl
start hawk
If you want Hawk2 to start automatically at boot time, execute the following command:
root #
systemctl
enable hawk
The Hawk2 Web interface uses the HTTPS protocol and port
7630
.
Instead of logging in to an individual cluster node with Hawk2, you can
configure a floating, virtual IP address (IPaddr
or
IPaddr2
) as a cluster resource. It does not need any
special configuration. It allows clients to connect to the Hawk service no
matter which physical node the service is running on.
When setting up the cluster with the
ha-cluster-bootstrap
scripts,
you will be asked whether to configure a virtual IP for cluster
administration.
On any machine, start a Web browser and enter the following URL:
https://HAWKSERVER:7630/
Replace HAWKSERVER with the IP address or host name of any cluster node running the Hawk Web service. If a virtual IP address has been configured for cluster administration with Hawk2, replace HAWKSERVER with the virtual IP address.
If a certificate warning appears when you try to access the URL for the first time, a self-signed certificate is in use. Self-signed certificates are not considered trustworthy by default.
To verify the certificate, ask your cluster operator for the certificate details.
To proceed anyway, you can add an exception in the browser to bypass the warning.
For information on how to replace the self-signed certificate with a certificate signed by an official Certificate Authority, refer to Replacing the Self-Signed Certificate.
On the Hawk2 login screen, enter the hacluster
user (or of any other
user that is a member of the
haclient
group).
Click
.After logging in to Hawk2, you will see a navigation bar on the left-hand side and a top-level row with several links on the right-hand side.
By default, users logged in as root
or
hacluster
have full
read-write access to all cluster configuration tasks. However,
Access Control Lists (ACLs) can be used to
define fine-grained access permissions.
If ACLs are enabled in the CRM, the available functions in Hawk2 depend
on the user role and their assigned access permissions. The
hacluster
.
Hawk2's top-level row shows the following entries:
: Allows you to set preferences for Hawk2 (for example, the language for the Web interface, or whether to display a warning if STONITH is disabled).
SUSE Linux Enterprise High Availability Extension documentation, read the release notes or report a bug.
: Access the: Click to log out.
Global cluster options control how the cluster behaves when confronted with certain situations. They are grouped into sets and can be viewed and modified with cluster management tools like Hawk2 and crmsh. The predefined values can usually be kept. However, to ensure the key functions of your cluster work correctly, you need to adjust the following parameters after basic cluster setup:
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› .The
screen opens. It displays the global cluster options and their current values.To display a short description of the parameter on the right-hand side of the screen, hover your mouse over a parameter.
Check the values for
and and adjust them, if necessary:
Set Section 6.2.2, “Global Option no-quorum-policy
” for more details.
If you need to disable fencing for any reason, set
no
. By default,
it is set to true
, because using STONITH devices is
necessary for normal cluster operation. According to the
default value, the cluster will refuse to start any resources if no
STONITH resources have been configured.
You must have a node fencing mechanism for your cluster.
The global cluster options
stonith-enabled
and
startup-fencing
must be set to
true
.
When you change them, you lose support.
To remove a parameter from the cluster configuration, click the
icon next to the parameter. If a parameter is deleted, the cluster will behave as if that parameter had the default value.To add a new parameter to the cluster configuration, choose one from the drop-down box.
If you need to change
or , proceed as follows:To adjust a value, either select a different value from the drop-down box or edit the value directly.
To add a new resource default or operation default, choose one from the empty drop-down box and enter a value. If there are default values, Hawk2 proposes them automatically.
To remove a parameter, click the Section 6.3.6, “Resource Options (Meta Attributes)” and Section 6.3.8, “Resource Operations”.
icon next to it. If no values are specified for and , the cluster uses the default values that are documented inConfirm your changes.
A cluster administrator needs to create cluster resources for every resource or application that runs on the servers in your cluster. Cluster resources can include Web sites, mail servers, databases, file systems, virtual machines, and any other server-based applications or services you want to make available to users at all times.
For an overview of the resource types you can create, refer to Section 6.3.3, “Types of Resources”. After you have specified the resource basics (ID, class, provider, and type), Hawk2 shows the following categories:
Determines which instance of a service the resource controls. For more information, refer to Section 6.3.7, “Instance Attributes (Parameters)”.
When creating a resource, Hawk2 automatically shows any required parameters. Edit them to get a valid resource configuration.
Needed for resource monitoring. For more information, refer to Section 6.3.8, “Resource Operations”.
When creating a resource, Hawk2 displays the most important resource
operations (monitor
, start
, and
stop
).
Tells the CRM how to treat a specific resource. For more information, refer to Section 6.3.6, “Resource Options (Meta Attributes)”.
When creating a resource, Hawk2 automatically lists the important meta
attributes for that resource (for example, the
target-role
attribute that defines the initial state of
a resource. By default, it is set to Stopped
, so the
resource will not start immediately).
Tells the CRM what capacity a certain resource requires from a node. For more information, refer to Section 7.6.8, “Configuring Placement of Resources Based on Load Impact”.
You can adjust the entries and values in those categories either during resource creation or later.
Sometimes a cluster administrator needs to know the cluster configuration. Hawk2 can show the current configuration in crm shell syntax, as XML and as a graph. To view the cluster configuration in crm shell syntax, from the left navigation bar select
› and click . To show the configuration in raw XML instead, click . Click for a graphical representation of the nodes and resources configured in the CIB. It also shows the relationships between resources.The Hawk2 wizard is a convenient way of setting up simple resources like a virtual IP address or an SBD STONITH resource, for example. It is also useful for complex configurations that include multiple resources, like the resource configuration for a DRBD block device or an Apache Web server. The wizard guides you through the configuration steps and provides information about the parameters you need to enter.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› .Expand the individual categories by clicking the arrow down icon next to them and select the desired wizard.
Follow the instructions on the screen. After the last configuration step,
the values you have entered.
Hawk2 shows which actions it is going to perform and what the
configuration looks like. Depending on the configuration, you might be
prompted for the root
password before you can
the configuration.
To create the most basic type of resource, proceed as follows:
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.In case a resource template exists on which you want to base the resource configuration, select the respective Procedure 7.6, “Adding a Resource Template”.
. For details about configuring templates, see
Select the resource agent lsb
, ocf
,
service
, stonith
, or
systemd
. For more information, see
Section 6.3.2, “Supported Resource Agent Classes”.
If you selected ocf
as class, specify the
of your OCF resource agent. The OCF
specification allows multiple vendors to supply the same resource agent.
From the
list, select the resource agent you want to use (for example, or ). A short description for this resource agent is displayed.With that, you have specified the resource basics.
The selection you get in the
list depends on the (and for OCF resources also on the ) you have chosen.To keep the
, , and as suggested by Hawk2, click to finish the configuration. A message at the top of the screen shows if the action has been successful.To adjust the parameters, operations, or meta attributes, refer to Section 7.5.5, “Modifying Resources”. To configure attributes for the resource, see Procedure 7.21, “Configuring the Capacity a Resource Requires”.
To create lots of resources with similar configurations, defining a resource template is the easiest way. After being defined, it can be referenced in primitives or in certain types of constraints. For detailed information about function and use of resource templates, refer to Section 6.5.3, “Resource Templates and Constraints”.
Resource templates are configured like primitive resources.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.Follow the instructions in Procedure 7.5, “Adding a Primitive Resource”, starting from Step 5.
If you have created a resource, you can edit its configuration at any time by adjusting parameters, operations, or meta attributes as needed.
Log in to Hawk2:
https://HAWKSERVER:7630/
On the Hawk2
screen, go to the list.In the
column, click the arrow down icon next to the resource or group you want to modify and select .The resource configuration screen opens.
To add a new parameter, operation, or meta attribute, select an entry from the empty drop-down box.
To edit any values in the
category, click the icon of the respective entry, enter a different value for the operation, and click .When you are finished, click the
button in the resource configuration screen to confirm your changes to the parameters, operations, or meta attributes.A message at the top of the screen shows if the action has been successful.
You must have a node fencing mechanism for your cluster.
The global cluster options
stonith-enabled
and
startup-fencing
must be set to
true
.
When you change them, you lose support.
By default, the global cluster option stonith-enabled
is
set to true
. If no STONITH resources have been defined,
the cluster will refuse to start any resources. Configure one or more
STONITH resources to complete the STONITH setup. To add a STONITH
resource for SBD, for libvirt (KVM/Xen) or for vCenter/ESX Server, the
easiest way is to use the Hawk2 wizard (see
Section 7.5.2, “Adding Resources with the Wizard”). While STONITH resources
are configured similarly to other resources, their behavior is different in
some respects. For details refer to Section 10.3, “STONITH Resources and Configuration”.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.From the
list, select the resource agent class .From the
list, select the STONITH plug-in to control your STONITH device. A short description for this plug-in is displayed.Hawk2 automatically shows the required
for the resource. Enter values for each parameter.Hawk2 displays the most important resource
and proposes default values. If you do not modify any settings here, Hawk2 adds the proposed operations and their default values when you confirm.If there is no reason to change them, keep the default
settings.Confirm your changes to create the STONITH resource.
A message at the top of the screen shows if the action has been successful.
To complete your fencing configuration, add constraints. For more details, refer to Chapter 10, Fencing and STONITH.
Some cluster resources depend on other components or resources. They require that each component or resource starts in a specific order and runs on the same server. To simplify this configuration SUSE Linux Enterprise High Availability Extension supports the concept of groups.
Resource groups contain a set of resources that need to be located together, be started sequentially and stopped in the reverse order. For an example of a resource group and more information about groups and their properties, refer to Section 6.3.5.1, “Groups”.
Groups must contain at least one resource, otherwise the configuration is not valid. While creating a group, Hawk2 allows you to create more primitives and add them to the group. For details, see Section 7.7.1, “Editing Resources and Groups”.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.To define the group members, select one or multiple entries in the list of “handle” icon on the right.
. Re-sort group members by dragging and dropping them into the order you want by using theIf needed, modify or add
.Click
to finish the configuration. A message at the top of the screen shows if the action has been successful.
If you want certain resources to run simultaneously on multiple nodes in
your cluster, configure these resources as clones. An example of a resource
that can be configured as a clone is
ocf:pacemaker:controld
for cluster file systems like
OCFS2. Any regular resources or resource groups can be cloned. Instances of
cloned resources may behave identically. However, they may also be
configured differently, depending on which node they are hosted on.
For an overview of the available types of resource clones, refer to Section 6.3.5.2, “Clones”.
Clones can either contain a primitive or a group as child resources. In Hawk2, child resources cannot be created or modified while creating a clone. Before adding a clone, create child resources and configure them as desired. For details, refer to Section 7.5.3, “Adding Simple Resources” or Section 7.5.7, “Adding Cluster Resource Groups”.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.From the
list, select the primitive or group to use as a sub-resource for the clone.If needed, modify or add
.Click
to finish the configuration. A message at the top of the screen shows if the action has been successful.
Multi-state resources are a specialization of clones. They allow the
instances to be in one of two operating modes (called
active/passive
, primary/secondary
, or
master/slave
). Multi-state resources must contain exactly
one group or one regular resource.
When configuring resource monitoring or constraints, multi-state resources have different requirements than simple resources. For details, see Pacemaker Explained, available from http://www.clusterlabs.org/pacemaker/doc/. Refer to section Multi-state - Resources That Have Multiple Modes.
Multi-state resources can either contain a primitive or a group as child resources. In Hawk2, child resources cannot be created or modified while creating a multi-state resource. Before adding a multi-state resource, create child resources and configure them as desired. For details, refer to Section 7.5.3, “Adding Simple Resources” or Section 7.5.7, “Adding Cluster Resource Groups”.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.From the
list, select the primitive or group to use as a sub-resource for the multi-state resource.If needed, modify or add
.Click
to finish the configuration. A message at the top of the screen shows if the action has been successful.
Tags are a way to refer to multiple resources at once, without creating any
colocation or ordering relationship between them. You can use tags for
grouping conceptually related resources. For example, if you have several
resources related to a database, you can add all related resources to a tag
named database
.
All resources belonging to a tag can be started or stopped with a single command.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.From the
list, select the resources you want to refer to with the tag.Click
to finish the configuration. A message at the top of the screen shows if the action has been successful.
The High Availability Extension does not only detect node failures, but also when an individual
resource on a node has failed. If you want to ensure that a resource is
running, configure resource monitoring for it. Usually, resources are only
monitored by the cluster while they are running. However, to detect
concurrency violations, also configure monitoring for resources which are
stopped. For resource monitoring, specify a timeout and/or start delay
value, and an interval. The interval tells the CRM how often it should check
the resource status. You can also set particular parameters such as
timeout
for start
or
stop
operations.
Log in to Hawk2:
https://HAWKSERVER:7630/
Add a resource as described in Procedure 7.5, “Adding a Primitive Resource” or select an existing primitive to edit.
Hawk2 automatically shows the most important
start
,
stop
, monitor
) and proposes default
values.
To see the attributes belonging to each proposed value, hover the mouse pointer over the respective value.
To change the suggested timeout
values for the
start
or stop
operation:
Click the pen icon next to the operation.
In the dialog that opens, enter a different value for the
timeout
parameter, for example 10
,
and confirm your change.
To change the suggested monitor
operation:
Click the pen icon next to the operation.
In the dialog that opens, enter a different value for the monitoring
interval
.
To configure resource monitoring in the case that the resource is stopped:
Select the role
entry from the empty drop-down box
below.
From the role
drop-down box, select
Stopped
.
Click
to confirm your changes and to close the dialog for the operation.Confirm your changes in the resource configuration screen. A message at the top of the screen shows if the action has been successful.
For the processes that take place if the resource monitor detects a failure, refer to Section 6.4, “Resource Monitoring”.
To view resource failures, switch to the
screen in Hawk2 and select the resource you are interested in. In the column click the arrow down icon and select . The dialog that opens lists recent actions performed for the resource. Failures are displayed in red. To view the resource details, click the magnifier icon in the column.After you have configured all resources, specify how the cluster should handle them correctly. Resource constraints let you specify on which cluster nodes resources can run, in which order to load resources, and what other resources a specific resource depends on.
For an overview of available types of constraints, refer to Section 6.5.1, “Types of Constraints”. When defining constraints, you also need to specify scores. For more information on scores and their implications in the cluster, see Section 6.5.2, “Scores and Infinity”.
A location constraint determines on which node a resource may be run, is preferably run, or may not be run. An example of a location constraint is to place all resources related to a certain database on the same node.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.From the list of
select the resource or resources for which to define the constraint.Enter a
. The score indicates the value you are assigning to this resource constraint. Positive values indicate the resource can run on the you specify in the next step. Negative values mean it should not run on that node. Constraints with higher scores are applied before those with lower scores.Some often-used values can also be set via the drop-down box:
To force the resources to run on the node, click the arrow
icon and select Always
. This sets the score to
INFINITY
.
If you never want the resources to run on the node, click the arrow icon
and select Never
. This sets the score to
-INFINITY
, meaning that the resources must not run on
the node.
To set the score to 0
, click the arrow icon and
select Advisory
. This disables the constraint. This
is useful when you want to set resource discovery but do not want to
constrain the resources.
Select a
.Click
to finish the configuration. A message at the top of the screen shows if the action has been successful.A colocational constraint tells the cluster which resources may or may not run together on a node. As a colocation constraint defines a dependency between resources, you need at least two resources to create a colocation constraint.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.Enter a
. The score determines the location relationship between the resources. Positive values indicate that the resources should run on the same node. Negative values indicate that the resources should not run on the same node. The score will be combined with other factors to decide where to put the resource.Some often-used values can also be set via the drop-down box:
To force the resources to run on the same node, click the
arrow icon and select Always
. This sets the score to
INFINITY
.
If you never want the resources to run on the same node, click the arrow
icon and select Never
. This sets the score to
-INFINITY
, meaning that the resources must not run on
the same node.
To define the resources for the constraint:
From the drop-down box in the
category, select a resource (or a template).The resource is added and a new empty drop-down box appears beneath.
Repeat this step to add more resources.
As the topmost resource depends on the next resource and so on, the cluster will first decide where to put the last resource, then place the depending ones based on that decision. If the constraint cannot be satisfied, the cluster may not allow the dependent resource to run.
To swap the order of resources within the colocation constraint, click the arrow up icon next to a resource to swap it with the entry above.
If needed, specify further parameters for each resource (such as
Started
, Stopped
,
Master
, Slave
,
Promote
, Demote
): Click the empty
drop-down box next to the resource and select the desired entry.
Click
to finish the configuration. A message at the top of the screen shows if the action has been successful.Order constraints define the order in which resources are started and stopped. As an order constraint defines a dependency between resources, you need at least two resources to create an order constraint.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› › .Enter a unique
.Enter a
. If the score is greater than zero, the order constraint is mandatory, otherwise it is optional.Some often-used values can also be set via the drop-down box:
If you want to make the order constraint mandatory, click the arrow icon
and select Mandatory
.
If you want the order constraint to be a suggestion only, click the
arrow icon and select Optional
.
Serialize
: To ensure that no two stop/start actions
occur concurrently for the resources, click the arrow icon and select
Serialize
. This makes sure that one resource must
complete starting before the other can be started. A typical use case
are resources that put a high load on the host during start-up.
For order constraints, you can usually keep the option
enabled. This specifies that resources are stopped in reverse order.To define the resources for the constraint:
From the drop-down box in the
category, select a resource (or a template).The resource is added and a new empty drop-down box appears beneath.
Repeat this step to add more resources.
The topmost resource will start first, then the second, etc. Usually the resources will be stopped in reverse order.
To swap the order of resources within the order constraint, click the arrow up icon next to a resource to swap it with the entry above.
If needed, specify further parameters for each resource (like
Started
, Stopped
,
Master
, Slave
,
Promote
, Demote
): Click the empty
drop-down box next to the resource and select the desired entry.
Confirm your changes to finish the configuration. A message at the top of the screen shows if the action has been successful.
As an alternative format for defining constraints, you can use Resource Sets. They have the same ordering semantics as Groups.
To use a resource set within a location constraint:
Proceed as outlined in Procedure 7.14, “Adding a Location Constraint”, apart from Step 4: Instead of selecting a single resource, select multiple resources by pressing Ctrl or Shift and mouse click. This creates a resource set within the location constraint.
To remove a resource from the location constraint, press Ctrl and click the resource again to deselect it.
To use a resource set within a colocation or order constraint:
Proceed as described in Procedure 7.15, “Adding a Colocation Constraint” or Procedure 7.16, “Adding an Order Constraint”, apart from the step where you define the resources for the constraint (Step 5.a or Step 6.a):
Add multiple resources.
To create a resource set, click the chain icon next to a resource to link it to the resource above. A resource set is visualized by a frame around the resources belonging to a set.
You can combine multiple resources in a resource set or create multiple resource sets.
To unlink a resource from the resource above, click the scissors icon next to the resource.
Confirm your changes to finish the constraint configuration.
For more information on configuring constraints and detailed background information about the basic concepts of ordering and colocation, refer to the documentation available at http://www.clusterlabs.org/pacemaker/doc/:
Pacemaker Explained, chapter Resource Constraints
Colocation Explained
Ordering Explained
A resource will be automatically restarted if it fails. If that cannot be
achieved on the current node, or it fails N times
on the current node, it will try to fail over to another node. You can
define several failures for resources (a
migration-threshold
), after which they will migrate to a
new node. If you have more than two nodes in your cluster, the node to which
a particular resource fails over is chosen by the High Availability software.
You can specify a specific node to which a resource will fail over by proceeding as follows:
Log in to Hawk2:
https://HAWKSERVER:7630/
Configure a location constraint for the resource as described in Procedure 7.14, “Adding a Location Constraint”.
Add the migration-threshold
meta attribute to the
resource as described in
Procedure 7.7: Modifying Parameters, Operations, or Meta Attributes for a Resource,
Step 4
and enter a value for the migration-threshold. The value should be
positive and less than INFINITY.
If you want to automatically expire the failcount for a resource, add the
failure-timeout
meta attribute to the resource as
described in
Procedure 7.5: Adding a Primitive Resource,
Step 4
and enter a for the
failure-timeout
.
If you want to specify additional failover nodes with preferences for a resource, create additional location constraints.
The process flow regarding migration thresholds and failcounts is demonstrated in Example 6.8, “Migration Threshold—Process Flow”.
Instead of letting the failcount for a resource expire automatically, you can also clean up failcounts for a resource manually at any time. Refer to Section 7.7.3, “Cleaning Up Resources” for details.
A resource may fail back to its original node when that node is back online and in the cluster. To prevent this or to specify a different node for the resource to fail back to, change the stickiness value of the resource. You can either specify the resource stickiness when creating it or afterward.
For the implications of different resource stickiness values, refer to Section 6.5.5, “Failback Nodes”.
Log in to Hawk2:
https://HAWKSERVER:7630/
Add the resource-stickiness
meta attribute to the
resource as described in
Procedure 7.7: Modifying Parameters, Operations, or Meta Attributes for a Resource,
Step 4.
Specify a value between -INFINITY
and
INFINITY
for resource-stickiness
.
Not all resources are equal. Some, such as Xen guests, require that the node hosting them meets their capacity requirements. If resources are placed so that their combined needs exceed the provided capacity, the performance of the resources diminishes or they fail.
To take this into account, the High Availability Extension allows you to specify the following parameters:
The capacity a certain node provides.
The capacity a certain resource requires.
An overall strategy for placement of resources.
For more details and a configuration example, refer to Section 6.5.6, “Placing Resources Based on Their Load Impact”.
Utilization attributes are used to configure both the resource's requirements and the capacity a node provides. You first need to configure a node's capacity before you can configure the capacity a resource requires.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› .On the
tab, select the node whose capacity you want to configure.In the
column, click the arrow down icon and select .The
screen opens.Below
, enter a name for a utilization attribute into the empty drop-down box.
The name can be arbitrary (for example, RAM_in_GB
).
Click the
icon to add the attribute.In the empty text box next to the attribute, enter an attribute value. The value must be an integer.
Add as many utilization attributes as you need and add values for all of them.
Confirm your changes. A message at the top of the screen shows if the action has been successful.
Configure the capacity a certain resource requires from a node either when creating a primitive resource or when editing an existing primitive resource.
Before you can add utilization attributes to a resource, you need to have set utilization attributes for your cluster nodes as described in Procedure 7.20.
Log in to Hawk2:
https://HAWKSERVER:7630/
To add a utilization attribute to an existing resource: Go to Section 7.7.1, “Editing Resources and Groups”.
› and open the resource configuration dialog as described inIf you create a new resource: Go to Section 7.5.3, “Adding Simple Resources”.
› and proceed as described inIn the resource configuration dialog, go to the
category.From the empty drop-down box, select one of the utilization attributes that you have configured for the nodes in Procedure 7.20.
In the empty text box next to the attribute, enter an attribute value. The value must be an integer.
Add as many utilization attributes as you need and add values for all of them.
Confirm your changes. A message at the top of the screen shows if the action has been successful.
After you have configured the capacities your nodes provide and the capacities your resources require, set the placement strategy in the global cluster options. Otherwise the capacity configurations have no effect. Several strategies are available to schedule the load: for example, you can concentrate it on as few nodes as possible, or balance it evenly over all available nodes. For more information, refer to Section 6.5.6, “Placing Resources Based on Their Load Impact”.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› to open the respective screen. It shows global cluster options and resource and operation defaults.
From the empty drop-down box in the upper part of the screen, select
placement-strategy
.
By default, its value is set to
, which means that utilization attributes and values are not considered.Depending on your requirements, set
to the appropriate value.Confirm your changes.
In addition to configuring your cluster resources, Hawk2 allows you to manage existing resources from the Section 7.8.1, “Monitoring a Single Cluster”.
screen. For a general overview of the screen refer toIn case you need to edit existing resources, go to the
screen. In the column, click the arrow down icon next to the resource or group you want to modify and select .The editing screen appears. If you edit a primitive resource, the following operations are available:
Copying the resource.
Renaming the resource (changing its ID).
Deleting the resource.
If you edit a group, the following operations are available:
Creating a new primitive which will be added to this group.
Renaming the group (changing its ID).
Re-sort group members by dragging and dropping them into the order you want using the “handle” icon on the right.
Before you start a cluster resource, make sure it is set up correctly. For example, if you use an Apache server as a cluster resource, set up the Apache server first. Complete the Apache configuration before starting the respective resource in your cluster.
When managing a resource via the High Availability Extension, the resource must not be started or stopped otherwise (outside of the cluster, for example manually or on boot or reboot). The High Availability Extension software is responsible for all service start or stop actions.
However, if you want to check if the service is configured properly, start it manually, but make sure that it is stopped again before the High Availability Extension takes over.
For interventions in resources that are currently managed by the cluster,
set the resource to maintenance mode
first. For details,
see Procedure 16.5, “Putting a Resource into Maintenance Mode with Hawk2”.
When creating a resource with Hawk2, you can set its initial state with
the target-role
meta attribute. If you set its value to
stopped
, the resource does not start automatically after
being created.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› . The list of also shows the .Select the resource to start. In its
column click the icon. To continue, confirm the message that appears.When the resource has started, Hawk2 changes the resource's
to green and shows on which node it is running.A resource will be automatically restarted if it fails, but each failure increases the resource's failcount.
If a migration-threshold
has been set for the resource,
the node will no longer run the resource when the number of failures reaches
the migration threshold.
A resource's failcount can either be reset automatically (by setting a
failure-timeout
option for the resource) or it can be
reset manually as described below.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
. The list of also shows the .Go to the resource to clean up. In the
column click the arrow down button and select . To continue, confirm the message that appears.
This executes the command crm resource cleanup
and
cleans up the resource on all nodes.
If you need to remove a resource from the cluster, follow the procedure below to avoid configuration errors:
Log in to Hawk2:
https://HAWKSERVER:7630/
Clean up the resource on all nodes as described in Procedure 7.24, “Cleaning Up A Resource”.
Stop the resource:
From the left navigation bar, select
› . The list of also shows the .In the
column click the button next to the resource.To continue, confirm the message that appears.
The
column will reflect the change when the resource is stopped.Delete the resource:
From the left navigation bar, select
› .In the list of
, go to the respective resource. From the column click the icon next to the resource.To continue, confirm the message that appears.
As mentioned in Section 7.6.6, “Specifying Resource Failover Nodes”, the cluster will fail over (migrate) resources automatically in case of software or hardware failures—according to certain parameters you can define (for example, migration threshold or resource stickiness). You can also manually migrate a resource to another node in the cluster. Or you decide to move the resource away from the current node and let the cluster decide where to put it.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› . The list of also shows the .In the list of
, select the respective resource.In the
column click the arrow down button and select .In the window that opens you have the following choices:
-INFINITY
score for the current
node.
Alternatively, you can move the resource to another node. This creates a
location constraint with an INFINITY
score for the
destination node.
Confirm your choice.
To allow a resource to move back again, proceed as follows:
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› . The list of also shows the .In the list of
, go to the respective resource.In the
column click the arrow down button and select . To continue, confirm the message that appears.
Hawk2 uses the crm_resource
--clear
command. The resource can move back to its original location or it may
stay where it is (depending on resource stickiness).
For more information, see Pacemaker Explained, available from http://www.clusterlabs.org/pacemaker/doc/. Refer to section Resource Migration.
Hawk2 has different screens for monitoring single clusters and multiple clusters: the
and the screen.To monitor a single cluster, use the
screen. After you have logged in to Hawk2, the screen is displayed by default. An icon in the upper right corner shows the cluster status at a glance. For further details, have a look at the following categories:If errors have occurred, they are shown at the top of the page.
Shows the configured resources including their
, (ID), (node on which they are running), and resource agent . From the column, you can start or stop a resource, trigger several actions, or view details. Actions that can be triggered include setting the resource to maintenance mode (or removing maintenance mode), migrating it to a different node, cleaning up the resource, showing any recent events, or editing the resource.
Shows the nodes belonging to the cluster site you are logged in to,
including the nodes' maintenance
or standby
flag for a
node. The column allows you to view recent
events for the node or further
details: for example, if a utilization
,
standby
or maintenance
attribute is
set for the respective node.
Only shown if tickets have been configured (for use with Geo clustering).
To monitor multiple clusters, use the Hawk2 Section D.2, “Configuring a Passwordless SSH Account”. However, the machine running Hawk2 does not even need to be part of any cluster for that purpose—it can be a separate, unrelated system.
. The cluster information displayed in the screen is stored on the server side. It is synchronized between the cluster nodes (if passwordless SSH access between the cluster nodes has been configured). For details, seeIn addition to the general Hawk2 Requirements, the following prerequisites need to be fulfilled to monitor multiple clusters with Hawk2:
All clusters to be monitored from Hawk2's SUSE Linux Enterprise High Availability Extension 15 SP1.
must be runningIf you did not replace the self-signed certificate for Hawk2 on every cluster node with your own certificate (or a certificate signed by an official Certificate Authority) yet, do the following: Log in to Hawk2 on every node in every cluster at least once. Verify the certificate (or add an exception in the browser to bypass the warning). Otherwise Hawk2 cannot connect to the cluster.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› .Hawk2 shows an overview of the resources and nodes on the current cluster site. In addition, it shows any
that have been configured for use with a Geo cluster. If you need information about the icons used in this view, click . To search for a resource ID, enter the name (ID) into the text box. To only show specific nodes, click the filter icon and select a filtering option.amsterdam
) #To add dashboards for multiple clusters:
Click
.
Enter the berlin
.
Enter the fully qualified host name of one of the nodes in the second
cluster. For example, charlie
.
Click
. Hawk2 will display a second tab for the newly added cluster site with an overview of its nodes and resources.If instead you are prompted to log in to this node by entering a
password, you probably did not connect to this node yet and have not
replaced the self-signed certificate. In that case, even after entering
the password, the connection will fail with the following message:
Error connecting to server. Retrying every 5 seconds... '
.
To proceed, see Replacing the Self-Signed Certificate.
To view more details for a cluster site or to manage it, switch to the site's tab and click the chain icon.
Hawk2 opens the
view for this site in a new browser window or tab. From there, you can administer this part of the Geo cluster.
To remove a cluster from the dashboard, click the x
icon on the right-hand side of the cluster's details.
Hawk2 provides a cluster simulator. It can be used for the following:
, including aStaging changes to the cluster and applying them as a single transaction, instead of having each change take effect immediately.
Simulating changes and cluster events, for example, to explore potential failure scenarios.
For example, batch mode can be used when creating groups of resources that depend on each other. Using batch mode, you can avoid applying intermediate or incomplete configurations to the cluster.
While batch mode is enabled, you can add or edit resources and constraints or change the cluster configuration. It is also possible to simulate events in the cluster, including nodes going online or offline, resource operations and tickets being granted or revoked. See Procedure 7.30, “Injecting Node, Resource or Ticket Events” for details.
The cluster simulator runs automatically after every change and shows the expected outcome in the user interface. For example, this also means: If you stop a resource while in batch mode, the user interface shows the resource as stopped—while actually, the resource is still running.
Some wizards include actions beyond mere cluster configuration. When using those wizards in batch mode, any changes that go beyond cluster configuration would be applied to the live system immediately.
Therefore wizards that require root
permission cannot be executed in
batch mode.
Log in to Hawk2:
https://HAWKSERVER:7630/
To activate the batch mode, select
from the top-level row.An additional bar appears below the top-level row. It indicates that batch mode is active and contains links to actions that you can execute in batch mode.
While batch mode is active, perform any changes to your cluster, like adding or editing resources and constraints or editing the cluster configuration.
The changes will be simulated and shown in all screens.
To view details of the changes you have made, select
from the batch mode bar. The window opens.
For any configuration changes it shows the difference between the live
state and the simulated changes in crmsh syntax: Lines starting with a
-
character represent the current state whereas lines
starting with +
show the proposed state.
To inject events or view even more details, see Procedure 7.30. Otherwise the window.
Choose to either
or the simulated changes and confirm your choice. This also deactivates batch mode and takes you back to normal mode.When running in batch mode, Hawk2 also allows you to inject
and .Let you change the state of a node. Available states are
, , and .
Let you change some properties of a resource. For example, you can set an
operation (like start
, stop
,
monitor
), the node it applies to, and the expected
result to be simulated.
Let you test the impact of granting and revoking tickets (used for Geo clusters).
Log in to Hawk2:
https://HAWKSERVER:7630/
If batch mode is not active yet, click
at the top-level row to switch to batch mode.In the batch mode bar, click
to open the window.To simulate a status change of a node:
Click
› .Select the
you want to manipulate and select its target .Confirm your changes. Your event is added to the queue of events listed in the
dialog.To simulate a resource operation:
Click
› .Select the
you want to manipulate and select the to simulate.If necessary, define an
.Select the
on which to run the operation and the targeted . Your event is added to the queue of events listed in the dialog.Confirm your changes.
To simulate a ticket action:
Click
› .Select the
you want to manipulate and select the to simulate.Confirm your changes. Your event is added to the queue of events listed in the
dialog.The Figure 7.18) shows a new line per injected event. Any event listed here is simulated immediately and is reflected on the screen.
dialog (If you have made any configuration changes, too, the difference between the live state and the simulated changes is shown below the injected events.
To remove an injected event, click the
icon next to it. Hawk2 updates the screen accordingly.To view more details about the simulation run, click
and choose one of the following:Shows a detailed summary.
shows the initial CIB state. shows what the CIB would look like after the transition.
Shows a graphical representation of the transition.
Shows an XML representation of the transition.
If you have reviewed the simulated changes, close the
window.To leave the batch mode, either
or the simulated changes.Hawk2 provides the following possibilities to view past events on the cluster (on different levels and in varying detail):
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› . It lists and .To view recent events of a resource:
Click
and select the respective resource.In the
column for the resource, click the arrow down button and select .Hawk2 opens a new window and displays a table view of the latest events.
To view recent events of a node:
Click
and select the respective node.In the
column for the node, select .Hawk2 opens a new window and displays a table view of the latest events.
From the left navigation bar, select
› to access the . The allows you to create detailed cluster reports and view transition information. It provides the following options:
Create a cluster report for a certain time. Hawk2 calls the
crm report
command to generate the report.
Allows you to upload crm report
archives that have
either been created with the crm shell directly or even on a different
cluster.
After reports have been generated or uploaded, they are shown below
. From the list of reports, you can show a report's details, download or delete the report.Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› .The
screen opens in the view. By default, the suggested time frame for a report is the last hour.To create a cluster report:
To immediately start a report, click
.To modify the time frame for the report, click anywhere on the suggested time frame and select another option from the drop-down box. You can also enter a
start date, end date and hour, respectively. To start the report, click .After the report has finished, it is shown below
.
To upload a cluster report, the crm report
archive must
be located on a file system that you can access with Hawk2. Proceed as
follows:
Switch to the
tab.for the cluster report archive and click .
After the report is uploaded, it is shown below
.To download or delete a report, click the respective icon next to the report in the
column.To view Report Details in History Explorer, click the report's name or select from the column.
Return to the list of reports by clicking the
button.Name of the report.
Start time of the report.
End time of the report.
Number of transitions plus time line of all transitions in the cluster that are covered by the report. To learn how to view more details for a transition, see Section 7.10.3.
Node events.
Resource events.
For each transition, the cluster saves a copy of the state which it provides
as input to pacemaker-schedulerd
.
The path to this archive is logged. All
pe-*
files are generated on the Designated
Coordinator (DC). As the DC can change in a cluster, there may be
pe-*
files from several nodes. Any pe-*
files are saved snapshots of the CIB, used as input of calculations by pacemaker-schedulerd
.
In Hawk2, you can display the name of each pe-*
file plus the time and node on which it was created. In addition, the
can visualize the following details,
based on the respective pe-*
file:
Shows snippets of logging data that belongs to the transition. Displays the output of the following command (including the resource agents' log messages):
crm history transition peinput
Shows the cluster configuration at the time that the
pe-*
file was created.
Shows the differences of configuration and status between the selected
pe-*
file and the following one.
Shows snippets of logging data that belongs to the transition. Displays the output of the following command:
crm history transition log peinput
This includes details from the following daemons:
pacemaker-schedulerd
,
pacemaker-controld
, and
pacemaker-execd
.
Shows a graphical representation of the transition. If you click
pacemaker-schedulerd
) and a graphical
visualization is generated.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› .If reports have already been generated or uploaded, they are shown in the list of Procedure 7.31.
. Otherwise generate or upload a report as described inClick the report's name or select Report Details in History Explorer.
from the column to open theTo access the transition details, you need to select a transition point in the transition time line that is shown below. Use the
and icons and the and icons to find the transition that you are interested in.
To display the name of a pe-input*
file plus the time
and node on which it was created, hover the mouse pointer over a
transition point in the time line.
To view the Transition Details in the History Explorer, click the transition point for which you want to know more.
To return to the list of reports, click the
button.
Hawk2 provides a wizard which checks and detects issues with your cluster.
After the analysis is complete, Hawk2 creates a cluster report with further
details. To verify cluster health and generate the report, Hawk2 requires
passwordless SSH access between the nodes. Otherwise it can only collect data
from the current node. If you have set up your cluster with the bootstrap scripts,
provided by the
ha-cluster-bootstrap
package, passwordless SSH access is already configured. In case you need to
configure it manually, see Section D.2, “Configuring a Passwordless SSH Account”.
Log in to Hawk2:
https://HAWKSERVER:7630/
From the left navigation bar, select
› .Expand the
category.Select the
wizard.Confirm with
.Enter the root password for your cluster and click
. Hawk2 will generate the report.To configure and manage cluster resources, either use the crm shell (crmsh) command line utility or Hawk2, a Web-based user interface.
This chapter introduces crm
, the command line tool
and covers an overview of this tool, how to use templates, and mainly
configuring and managing cluster resources: creating basic and advanced
types of resources (groups and clones), configuring constraints,
specifying failover nodes and failback nodes, configuring resource
monitoring, starting, cleaning up or removing resources, and migrating
resources manually.
Sufficient privileges are necessary to manage a cluster. The
crm
command and its subcommands need to be run either
as root
user or as the CRM owner user (typically the user
hacluster
).
However, the user
option allows you to run
crm
and its subcommands as a regular (unprivileged)
user and to change its ID using sudo
whenever
necessary. For example, with the following command crm
will use hacluster
as the
privileged user ID:
root #
crm
options user hacluster
Note that you need to set up /etc/sudoers
so that
sudo
does not ask for a password.
The crm
command has several subcommands which manage
resources, CIBs, nodes, resource agents, and others. It offers a thorough
help system with embedded examples. All examples follow a naming
convention described in
Appendix B.
By using crm without arguments (or with only one sublevel as argument), the crm shell enters the interactive mode. This mode is indicated by the following prompt:
crm(live/HOSTNAME)
For readability reasons, we omit the host name in the interactive crm prompts in our documentation. We only include the host name if you need to run the interactive shell on a specific node, like alice for example:
crm(live/alice)
Help can be accessed in several ways:
To output the usage of crm
and its command line
options:
root #
crm
--help
To give a list of all available commands:
root #
crm
help
To access other help sections, not just the command reference:
root #
crm
help topics
To view the extensive help text of the configure
subcommand:
root #
crm
configure help
To print the syntax, its usage, and examples of the group
subcommand of configure
:
root #
crm
configure help group
This is the same:
root #
crm
help configure group
Almost all output of the help
subcommand (do not mix
it up with the --help
option) opens a text viewer. This
text viewer allows you to scroll up or down and read the help text more
comfortably. To leave the text viewer, press the Q key.
The crmsh supports full tab completion in Bash directly, not only
for the interactive shell. For example, typing crm help
config
→| will complete the word
like in the interactive shell.
The crm
command itself can be used in the following
ways:
Directly:
Concatenate all subcommands to crm
, press
Enter and you see the output immediately. For
example, enter crm
help ra
to get
information about the ra
subcommand (resource
agents).
It is possible to abbreviate subcommands as long as they are
unique. For example, you can shorten status
as
st
and crmsh will know what you have meant.
Another feature is to shorten parameters. Usually, you add
parameters through the params
keyword.
You can leave out the params section if it is the first and only section.
For example, this line:
root #
crm
primitive ipaddr ocf:heartbeat:IPaddr2 params ip=192.168.0.55
is equivalent to this line:
root #
crm
primitive ipaddr ocf:heartbeat:IPaddr2 ip=192.168.0.55
As crm Shell Script:
Crm shell scripts contain subcommands of crm
.
For more information, see Section 8.1.4, “Using crmsh's Shell Scripts”.
As crmsh Cluster Scripts: These are a collection of metadata, references to RPM packages,
configuration files, and crmsh subcommands bundled under a single,
yet descriptive name. They are managed through the
crm script
command.
Do not confuse them with crmsh shell scripts: although both share some common objectives, the crm shell scripts only contain subcommands whereas cluster scripts incorporate much more than a simple enumeration of commands. For more information, see Section 8.1.5, “Using crmsh's Cluster Scripts”.
Interactive as Internal Shell:
Type crm
to enter the internal shell. The prompt
changes to crm(live)
. With
help
you can get an overview of the available
subcommands. As the internal shell has different levels of
subcommands, you can “enter” one by typing this
subcommand and press Enter.
For example, if you type resource
you enter the
resource management level. Your prompt changes to
crm(live)resource#
. To leave the
internal shell, use the commands quit
,
bye
, or exit
. If you need to go
one level back, use back
, up
,
end
, or cd
.
You can enter the level directly by typing crm
and
the respective subcommand(s) without any options and press
Enter.
The internal shell supports also tab completion for subcommands and
resources. Type the beginning of a command, press
→| and crm
completes the
respective object.
In addition to previously explained methods, crmsh also supports
synchronous command execution. Use the -w
option to
activate it. If you have started crm
without
-w
, you can enable it later with the user preference's
wait
set to yes
(options
wait yes
). If this option is enabled, crm
waits until the transition is finished. Whenever a transaction is
started, dots are printed to indicate progress. Synchronous command
execution is only applicable for commands like resource
start
.
The crm
tool has management capability (the
subcommands resource
and node
)
and can be used for configuration (cib
,
configure
).
The following subsections give you an overview of some important aspects
of the crm
tool.
As you need to deal with resource agents in your cluster configuration
all the time, the crm
tool contains the
ra
command. Use it to show information about resource
agents and to manage them (for additional information, see also
Section 6.3.2, “Supported Resource Agent Classes”):
root #
crm
racrm(live)ra#
The command classes
lists all classes and providers:
crm(live)ra#
classes
lsb ocf / heartbeat linbit lvm2 ocfs2 pacemaker service stonith systemd
To get an overview of all available resource agents for a class (and
provider) use the list
command:
crm(live)ra#
list
ocf AoEtarget AudibleAlarm CTDB ClusterMon Delay Dummy EvmsSCC Evmsd Filesystem HealthCPU HealthSMART ICP IPaddr IPaddr2 IPsrcaddr IPv6addr LVM LinuxSCSI MailTo ManageRAID ManageVE Pure-FTPd Raid1 Route SAPDatabase SAPInstance SendArp ServeRAID ...
An overview of a resource agent can be viewed with
info
:
crm(live)ra#
info
ocf:linbit:drbd This resource agent manages a DRBD* resource as a master/slave resource. DRBD is a shared-nothing replicated storage device. (ocf:linbit:drbd) Master/Slave OCF Resource Agent for DRBD Parameters (* denotes required, [] the default): drbd_resource* (string): drbd resource name The name of the drbd resource from the drbd.conf file. drbdconf (string, [/etc/drbd.conf]): Path to drbd.conf Full path to the drbd.conf file. Operations' defaults (advisory minimum): start timeout=240 promote timeout=90 demote timeout=90 notify timeout=90 stop timeout=100 monitor_Slave_0 interval=20 timeout=20 start-delay=1m monitor_Master_0 interval=10 timeout=20 start-delay=1m
Leave the viewer by pressing Q.
crm
Directly
In the former example we used the internal shell of the
crm
command. However, you do not necessarily need to
use it. You get the same results if you add the respective subcommands
to crm
. For example, you can list all the OCF
resource agents by entering crm
ra list
ocf
in your shell.
The crmsh shell scripts provide a convenient way to enumerate crmsh subcommands into a file. This makes it easy to comment specific lines or to replay them later. Keep in mind that a crmsh shell script can contain only crmsh subcommands. Any other commands are not allowed.
Before you can use a crmsh shell script, create a file with specific commands. For example, the following file prints the status of the cluster and gives a list of all nodes:
# A small example file with some crm subcommandsstatus
node
list
Any line starting with the hash symbol (#
) is a
comment and is ignored. If a line is too long, insert a backslash
(\
) at the end and continue in the next line. It is
recommended to indent lines that belong to a certain subcommand to improve
readability.
To use this script, use one of the following methods:
root #
crm
-f example.cliroot #
crm
< example.cli
Collecting information from all cluster nodes and deploying any changes is a key cluster administration task. Instead of performing the same procedures manually on different nodes (which is error-prone), you can use the crmsh cluster scripts.
Do not confuse them with the crmsh shell scripts, which are explained in Section 8.1.4, “Using crmsh's Shell Scripts”.
In contrast to crmsh shell scripts, cluster scripts performs additional tasks like:
Installing software that is required for a specific task.
Creating or modifying any configuration files.
Collecting information and reporting potential problems with the cluster.
Deploying the changes to all nodes.
crmsh cluster scripts do not replace other tools for managing clusters—they provide an integrated way to perform the above tasks across the cluster. Find detailed information at http://crmsh.github.io/scripts/.
To get a list of all available cluster scripts, run:
root #
crm
script list
To view the components of a script, use the
show
command and the name of the cluster script,
for example:
root #
crm
script show mailto mailto (Basic) MailTo This is a resource agent for MailTo. It sends email to a sysadmin whenever a takeover occurs. 1. Notifies recipients by email in the event of resource takeover id (required) (unique) Identifier for the cluster resource email (required) Email address subject Subject
The output of show
contains a title, a
short description, and a procedure. Each procedure is divided
into a series of steps, performed in the given order.
Each step contains a list of required and optional parameters, along with a short description and its default value.
Each cluster script understands a set of common parameters. These parameters can be passed to any script:
Parameter | Argument | Description |
---|---|---|
action | INDEX | If set, only execute a single action (index, as returned by verify) |
dry_run | BOOL | If set, simulate execution only (default: no) |
nodes | LIST | List of nodes to execute the script for |
port | NUMBER | Port to connect to |
statefile | FILE | When single-stepping, the state is saved in the given file |
sudo | BOOL | If set, crm will prompt for a sudo password and use sudo where appropriate (default: no) |
timeout | NUMBER | Execution timeout in seconds (default: 600) |
user | USER | Run script as the given user |
Before running a cluster script, review the actions that it will perform and verify its parameters to avoid problems. A cluster script can potentially perform a series of actions and may fail for various reasons. Thus, verifying your parameters before running it helps to avoid problems.
For example, the mailto
resource agent
requires a unique identifier and an e-mail address. To verify these
parameters, run:
root #
crm
script verify mailto id=sysadmin email=tux@example.org 1. Ensure mail package is installed mailx 2. Configure cluster resources primitive sysadmin ocf:heartbeat:MailTo email="tux@example.org" op start timeout="10" op stop timeout="10" op monitor interval="10" timeout="10" clone c-sysadmin sysadmin
The verify
prints the steps and replaces
any placeholders with your given parameters. If verify
finds any problems, it will report it.
If everything is ok, replace the verify
command with run
:
root #
crm
script run mailto id=sysadmin email=tux@example.org INFO: MailTo INFO: Nodes: alice, bob OK: Ensure mail package is installed OK: Configure cluster resources
Check whether your resource is integrated into your cluster
with crm status
:
root #
crm
status [...] Clone Set: c-sysadmin [sysadmin] Started: [ alice bob ]
The use of configuration templates is deprecated and will be removed in the future. Configuration templates will be replaced by cluster scripts, see Section 8.1.5, “Using crmsh's Cluster Scripts”.
Configuration templates are ready-made cluster configurations for crmsh. Do not confuse them with the resource templates (as described in Section 8.3.3, “Creating Resource Templates”). Those are templates for the cluster and not for the crm shell.
Configuration templates require minimum effort to be tailored to the particular user's needs. Whenever a template creates a configuration, warning messages give hints which can be edited later for further customization.
The following procedure shows how to create a simple yet functional Apache configuration:
Log in as root
and start the crm
interactive shell:
root #
crm
configure
Create a new configuration from a configuration template:
Switch to the template
subcommand:
crm(live)configure#
template
List the available configuration templates:
crm(live)configure template#
list
templates gfs2-base filesystem virtual-ip apache clvm ocfs2 gfs2
Decide which configuration template you need. As we need an Apache
configuration, we select the apache
template and
name it g-intranet
:
crm(live)configure template#
new
g-intranet apache INFO: pulling in template apache INFO: pulling in template virtual-ip
Define your parameters:
List the configuration you have created:
crm(live)configure template#
list
g-intranet
Display the minimum required changes that need to be filled out by you:
crm(live)configure template#
show
ERROR: 23: required parameter ip not set ERROR: 61: required parameter id not set ERROR: 65: required parameter configfile not set
Invoke your preferred text editor and fill out all lines that have been displayed as errors in Step 3.b:
crm(live)configure template#
edit
Show the configuration and check whether it is valid (bold text depends on the configuration you have entered in Step 3.c):
crm(live)configure template#
show
primitive virtual-ip ocf:heartbeat:IPaddr \ params ip="192.168.1.101" primitive apache ocf:heartbeat:apache \ params configfile="/etc/apache2/httpd.conf" monitor apache 120s:60s group g-intranet \ apache virtual-ip
Apply the configuration:
crm(live)configure template#
apply
crm(live)configure#
cd ..
crm(live)configure#
show
Submit your changes to the CIB:
crm(live)configure#
commit
It is possible to simplify the commands even more, if you know the details. The above procedure can be summarized with the following command on the shell:
root #
crm
configure template \ new g-intranet apache params \ configfile="/etc/apache2/httpd.conf" ip="192.168.1.101"
If you are inside your internal crm
shell, use the
following command:
crm(live)configure template#
new
intranet apache params \ configfile="/etc/apache2/httpd.conf" ip="192.168.1.101"
However, the previous command only creates its configuration from the configuration template. It does not apply nor commit it to the CIB.
A shadow configuration is used to test different configuration scenarios. If you have created several shadow configurations, you can test them one by one to see the effects of your changes.
The usual process looks like this:
Log in as root
and start the crm
interactive shell:
root #
crm
configure
Create a new shadow configuration:
crm(live)configure#
cib
new myNewConfig INFO: myNewConfig shadow CIB created
If you omit the name of the shadow CIB, a temporary name
@tmp@
is created.
To copy the current live configuration into your shadow configuration, use the following command, otherwise skip this step:
crm(myNewConfig)# cib
reset myNewConfig
The previous command makes it easier to modify any existing resources later.
Make your changes as usual. After you have created the shadow configuration, all changes go there. To save all your changes, use the following command:
crm(myNewConfig)# commit
If you need the live cluster configuration again, switch back with the following command:
crm(myNewConfig)configure#cib
use livecrm(live)#
Before loading your configuration changes back into the cluster, it is
recommended to review your changes with ptest
. The
ptest
command can show a diagram of actions that will
be induced by committing the changes. You need the
graphviz package to display the diagrams. The
following example is a transcript, adding a monitor operation:
root #
crm
configurecrm(live)configure#
show
fence-bob primitive fence-bob stonith:apcsmart \ params hostlist="bob"crm(live)configure#
monitor
fence-bob 120m:60scrm(live)configure#
show
changed primitive fence-bob stonith:apcsmart \ params hostlist="bob" \ op monitor interval="120m" timeout="60s"crm(live)configure#
ptestcrm(live)configure#
commit
To output a cluster diagram, use the command
crm
configure graph
. It displays
the current configuration on its current window, therefore requiring
X11.
If you prefer Scalable Vector Graphics (SVG), use the following command:
root #
crm
configure graph dot config.svg svg
Corosync is the underlying messaging layer for most HA clusters. The
corosync
subcommand provides commands for editing and
managing the Corosync configuration.
For example, to list the status of the cluster, use
status
:
root #
crm
corosync status Printing ring status. Local node ID 175704363 RING ID 0 id = 10.121.9.43 status = ring 0 active with no faults Quorum information ------------------ Date: Thu May 8 16:41:56 2014 Quorum provider: corosync_votequorum Nodes: 2 Node ID: 175704363 Ring ID: 4032 Quorate: Yes Votequorum information ---------------------- Expected votes: 2 Highest expected: 2 Total votes: 2 Quorum: 2 Flags: Quorate Membership information ---------------------- Nodeid Votes Name 175704363 1 alice.example.com (local) 175704619 1 bob.example.com
The diff
command is very helpful: It compares the
Corosync configuration on all nodes (if not stated otherwise) and
prints the difference between:
root #
crm
corosync diff --- bob +++ alice @@ -46,2 +46,2 @@ - expected_votes: 2 - two_node: 1 + expected_votes: 1 + two_node: 0
For more details, see http://crmsh.nongnu.org/crm.8.html#cmdhelp_corosync.
As a cluster administrator, you need to create cluster resources for every resource or application you run on servers in your cluster. Cluster resources can include Web sites, e-mail servers, databases, file systems, virtual machines, and any other server-based applications or services you want to make available to users at all times.
For an overview of resource types you can create, refer to Section 6.3.3, “Types of Resources”.
Parts or all of the configuration can be loaded from a local file or a network URL. Three different methods can be defined:
replace
This option replaces the current configuration with the new source configuration.
update
This option tries to import the source configuration. It adds new items or updates existing items to the current configuration.
push
This option imports the content from the source into the current
configuration (same as update
). However, it removes
objects that are not available in the new configuration.
To load the new configuration from the file mycluster-config.txt
use the following syntax:
root #
crm
configure load push mycluster-config.txt
There are three types of RAs (Resource Agents) available with the cluster (for background information, see Section 6.3.2, “Supported Resource Agent Classes”). To add a new resource to the cluster, proceed as follows:
Log in as root
and start the crm
tool:
root #
crm
configure
Configure a primitive IP address:
crm(live)configure#
primitive
myIP ocf:heartbeat:IPaddr \ params ip=127.0.0.99 op monitor interval=60s
The previous command configures a “primitive” with the
name myIP
. You need to choose a class (here
ocf
), provider (heartbeat
), and
type (IPaddr
). Furthermore, this primitive expects
other parameters like the IP address. Change the address to your
setup.
Display and review the changes you have made:
crm(live)configure#
show
Commit your changes to take effect:
crm(live)configure#
commit
If you want to create several resources with similar configurations, a
resource template simplifies the task. See also
Section 6.5.3, “Resource Templates and Constraints” for some
basic background information. Do not confuse them with the
“normal” templates from
Section 8.1.6, “Using Configuration Templates”. Use the
rsc_template
command to get familiar with the syntax:
root #
crm
configure rsc_template usage: rsc_template <name> [<class>:[<provider>:]]<type> [params <param>=<value> [<param>=<value>...]] [meta <attribute>=<value> [<attribute>=<value>...]] [utilization <attribute>=<value> [<attribute>=<value>...]] [operations id_spec [op op_type [<attribute>=<value>...] ...]]
For example, the following command creates a new resource template with
the name BigVM
derived from the
ocf:heartbeat:Xen
resource and some default values
and operations:
crm(live)configure#
rsc_template
BigVM ocf:heartbeat:Xen \ params allow_mem_management="true" \ op monitor timeout=60s interval=15s \ op stop timeout=10m \ op start timeout=10m
Once you defined the new resource template, you can use it in primitives
or reference it in order, colocation, or rsc_ticket constraints. To
reference the resource template, use the @
sign:
crm(live)configure#
primitive
MyVM1 @BigVM \ params xmfile="/etc/xen/shared-vm/MyVM1" name="MyVM1"
The new primitive MyVM1 is going to inherit everything from the BigVM resource templates. For example, the equivalent of the above two would be:
crm(live)configure#
primitive
MyVM1 ocf:heartbeat:Xen \ params xmfile="/etc/xen/shared-vm/MyVM1" name="MyVM1" \ params allow_mem_management="true" \ op monitor timeout=60s interval=15s \ op stop timeout=10m \ op start timeout=10m
If you want to overwrite some options or operations, add them to your (primitive) definition. For example, the following new primitive MyVM2 doubles the timeout for monitor operations but leaves others untouched:
crm(live)configure#
primitive
MyVM2 @BigVM \ params xmfile="/etc/xen/shared-vm/MyVM2" name="MyVM2" \ op monitor timeout=120s interval=30s
A resource template may be referenced in constraints to stand for all primitives which are derived from that template. This helps to produce a more concise and clear cluster configuration. Resource template references are allowed in all constraints except location constraints. Colocation constraints may not contain more than one template reference.
From the crm
perspective, a STONITH device is
just another resource. To create a STONITH resource, proceed as
follows:
Log in as root
and start the crm
interactive shell:
root #
crm
configure
Get a list of all STONITH types with the following command:
crm(live)#
ra
list stonith apcmaster apcmastersnmp apcsmart baytech bladehpi cyclades drac3 external/drac5 external/dracmc-telnet external/hetzner external/hmchttp external/ibmrsa external/ibmrsa-telnet external/ipmi external/ippower9258 external/kdumpcheck external/libvirt external/nut external/rackpdu external/riloe external/sbd external/vcenter external/vmware external/xen0 external/xen0-ha fence_legacy ibmhmc ipmilan meatware nw_rpc100s rcd_serial rps10 suicide wti_mpc wti_nps
Choose a STONITH type from the above list and view the list of possible options. Use the following command:
crm(live)#
ra
info stonith:external/ipmi IPMI STONITH external device (stonith:external/ipmi) ipmitool based power management. Apparently, the power off method of ipmitool is intercepted by ACPI which then makes a regular shutdown. If case of a split brain on a two-node it may happen that no node survives. For two-node clusters use only the reset method. Parameters (* denotes required, [] the default): hostname (string): Hostname The name of the host to be managed by this STONITH device. ...
Create the STONITH resource with the stonith
class, the type you have chosen in
Step 3,
and the respective parameters if needed, for example:
crm(live)#
configure
crm(live)configure#
primitive
my-stonith stonith:external/ipmi \ params hostname="alice" \ ipaddr="192.168.1.221" \ userid="admin" passwd="secret" \ op monitor interval=60m timeout=120s
Having all the resources configured is only one part of the job. Even if the cluster knows all needed resources, it might still not be able to handle them correctly. For example, try not to mount the file system on the slave node of DRBD (in fact, this would fail with DRBD). Define constraints to make these kind of information available to the cluster.
For more information about constraints, see Section 6.5, “Resource Constraints”.
The location
command defines on which nodes a
resource may be run, may not be run or is preferred to be run.
This type of constraint may be added multiple times for each resource.
All location
constraints are evaluated for a given
resource. A simple example that expresses a preference to run the
resource fs1
on the node with the name
alice
to 100 would be the
following:
crm(live)configure#
location
loc-fs1 fs1 100: alice
Another example is a location with pingd:
crm(live)configure#
primitive
pingd pingd \ params name=pingd dampen=5s multiplier=100 host_list="r1 r2"crm(live)configure#
location
loc-node_pref internal_www \ rule 50: #uname eq alice \ rule pingd: defined pingd
Another use case for location constraints are grouping primitives as a
resource set. This can be useful if several
resources depend on, for example, a ping attribute for network
connectivity. In former times, the -inf/ping
rules
needed to be duplicated several times in the configuration, making it
unnecessarily complex.
The following example creates a resource set
loc-alice
, referencing the virtual IP addresses
vip1
and vip2
:
crm(live)configure#
primitive
vip1 ocf:heartbeat:IPaddr2 params ip=192.168.1.5crm(live)configure#
primitive
vip2 ocf:heartbeat:IPaddr2 params ip=192.168.1.6crm(live)configure#
location
loc-alice { vip1 vip2 } inf: alice
In some cases it is much more efficient and convenient to use resource
patterns for your location
command. A resource
pattern is a regular expression between two slashes. For example, the
above virtual IP addresses can be all matched with the following:
crm(live)configure#
location
loc-alice /vip.*/ inf: alice
The colocation
command is used to define what
resources should run on the same or on different hosts.
It is only possible to set a score of either +inf or -inf, defining resources that must always or must never run on the same node. It is also possible to use non-infinite scores. In that case the colocation is called advisory and the cluster may decide not to follow them in favor of not stopping other resources if there is a conflict.
For example, to run the resources with the IDs
filesystem_resource
and nfs_group
always on the same host, use the following constraint:
crm(live)configure#
colocation
nfs_on_filesystem inf: nfs_group filesystem_resource
For a master slave configuration, it is necessary to know if the current node is a master in addition to running the resource locally.
Sometimes it is useful to be able to place a group of resources on the same node (defining a colocation constraint), but without having hard dependencies between the resources.
Use the command weak-bond
if you want to place
resources on the same node, but without any action if one of them
fails.
root #
crm
configure assist weak-bond RES1 RES2
The implementation of weak-bond
creates a dummy
resource and a colocation constraint with the given resources
automatically.
The order
command defines a sequence of action.
Sometimes it is necessary to provide an order of resource actions or operations. For example, you cannot mount a file system before the device is available to a system. Ordering constraints can be used to start or stop a service right before or after a different resource meets a special condition, such as being started, stopped, or promoted to master.
Use the following command in the crm
shell to
configure an ordering constraint:
crm(live)configure#
order
nfs_after_filesystem mandatory: filesystem_resource nfs_group
The example used for this section would not work without additional constraints. It is essential that all resources run on the same machine as the master of the DRBD resource. The DRBD resource must be master before any other resource starts. Trying to mount the DRBD device when it is not the master simply fails. The following constraints must be fulfilled:
The file system must always be on the same node as the master of the DRBD resource.
crm(live)configure#
colocation
filesystem_on_master inf: \ filesystem_resource drbd_resource:Master
The NFS server and the IP address must be on the same node as the file system.
crm(live)configure#
colocation
nfs_with_fs inf: \ nfs_group filesystem_resource
The NFS server and the IP address start after the file system is mounted:
crm(live)configure#
order
nfs_second mandatory: \ filesystem_resource:start nfs_group
The file system must be mounted on a node after the DRBD resource is promoted to master on this node.
crm(live)configure#
order
drbd_first inf: \ drbd_resource:promote filesystem_resource:start
To determine a resource failover, use the meta attribute migration-threshold. In case failcount exceeds migration-threshold on all nodes, the resource will remain stopped. For example:
crm(live)configure#
location
rsc1-alice rsc1 100: alice
Normally, rsc1 prefers to run on alice. If it fails there, migration-threshold is checked and compared to the failcount. If failcount >= migration-threshold then it is migrated to the node with the next best preference.
Start failures set the failcount to inf depend on the
start-failure-is-fatal
option. Stop failures cause
fencing. If there is no STONITH defined, the resource will not migrate.
For an overview, refer to Section 6.5.4, “Failover Nodes”.
A resource might fail back to its original node when that node is back online and in the cluster. To prevent a resource from failing back to the node that it was running on, or to specify a different node for the resource to fail back to, change its resource stickiness value. You can either specify resource stickiness when you are creating a resource or afterward.
For an overview, refer to Section 6.5.5, “Failback Nodes”.
Some resources may have specific capacity requirements such as minimum amount of memory. Otherwise, they may fail to start completely or run with degraded performance.
To take this into account, the High Availability Extension allows you to specify the following parameters:
The capacity a certain node provides.
The capacity a certain resource requires.
An overall strategy for placement of resources.
For detailed background information about the parameters and a configuration example, refer to Section 6.5.6, “Placing Resources Based on Their Load Impact”.
To configure the resource's requirements and the capacity a node
provides, use utilization attributes.
You can name the utilization attributes according to your preferences
and define as many name/value pairs as your configuration needs. In
certain cases, some agents update the utilization themselves, for
example the VirtualDomain
.
In the following example, we assume that you already have a basic configuration of cluster nodes and resources. You now additionally want to configure the capacities a certain node provides and the capacity a certain resource requires.
crm
#
Log in as root
and start the crm
interactive shell:
root #
crm
configure
To specify the capacity a node provides, use the following command and replace the placeholder NODE_1 with the name of your node:
crm(live)configure#
node
NODE_1 utilization memory=16384 cpu=8
With these values, NODE_1 would be assumed to provide 16GB of memory and 8 CPU cores to resources.
To specify the capacity a resource requires, use:
crm(live)configure#
primitive
xen1 ocf:heartbeat:Xen ... \ utilization memory=4096 cpu=4
This would make the resource consume 4096 of those memory units from NODE_1, and 4 of the CPU units.
Configure the placement strategy with the property
command:
crm(live)configure#
property
...
The following values are available:
default
(default value)Utilization values are not considered. Resources are allocated according to location scoring. If scores are equal, resources are evenly distributed across nodes.
utilization
Utilization values are considered when deciding if a node has enough free capacity to satisfy a resource's requirements. However, load-balancing is still done based on the number of resources allocated to a node.
minimal
Utilization values are considered when deciding if a node has enough free capacity to satisfy a resource's requirements. An attempt is made to concentrate the resources on as few nodes as possible (to achieve power savings on the remaining nodes).
balanced
Utilization values are considered when deciding if a node has enough free capacity to satisfy a resource's requirements. An attempt is made to distribute the resources evenly, thus optimizing resource performance.
The available placement strategies are best-effort—they do not yet use complex heuristic solvers to always reach optimum allocation results. Ensure that resource priorities are properly set so that your most important resources are scheduled first.
Commit your changes before leaving crmsh:
crm(live)configure#
commit
The following example demonstrates a three node cluster of equal nodes, with 4 virtual machines:
crm(live)configure#
node
alice utilization memory="4000"crm(live)configure#
node
bob utilization memory="4000"crm(live)configure#
node
charlie utilization memory="4000"crm(live)configure#
primitive
xenA ocf:heartbeat:Xen \ utilization hv_memory="3500" meta priority="10" \ params xmfile="/etc/xen/shared-vm/vm1"crm(live)configure#
primitive
xenB ocf:heartbeat:Xen \ utilization hv_memory="2000" meta priority="1" \ params xmfile="/etc/xen/shared-vm/vm2"crm(live)configure#
primitive
xenC ocf:heartbeat:Xen \ utilization hv_memory="2000" meta priority="1" \ params xmfile="/etc/xen/shared-vm/vm3"crm(live)configure#
primitive
xenD ocf:heartbeat:Xen \ utilization hv_memory="1000" meta priority="5" \ params xmfile="/etc/xen/shared-vm/vm4"crm(live)configure#
property
placement-strategy="minimal"
With all three nodes up, xenA will be placed onto a node first, followed by xenD. xenB and xenC would either be allocated together or one of them with xenD.
If one node failed, too little total memory would be available to host them all. xenA would be ensured to be allocated, as would xenD. However, only one of xenB or xenC could still be placed, and since their priority is equal, the result is not defined yet. To resolve this ambiguity as well, you would need to set a higher priority for either one.
To monitor a resource, there are two possibilities: either define a
monitor operation with the op
keyword or use the
monitor
command. The following example configures an
Apache resource and monitors it every 60 seconds with the
op
keyword:
crm(live)configure#
primitive
apache apache \ params ... \ op monitor interval=60s timeout=30s
The same can be done with:
crm(live)configure#
primitive
apache apache \ params ...crm(live)configure#
monitor
apache 60s:30s
For an overview, refer to Section 6.4, “Resource Monitoring”.
One of the most common elements of a cluster is a set of resources that needs to be located together. Start sequentially and stop in the reverse order. To simplify this configuration we support the concept of groups. The following example creates two primitives (an IP address and an e-mail resource):
Run the crm
command as system administrator. The
prompt changes to crm(live)
.
Configure the primitives:
crm(live)#
configure
crm(live)configure#
primitive
Public-IP ocf:heartbeat:IPaddr \ params ip=1.2.3.4 id= Public-IPcrm(live)configure#
primitive
Email systemd:postfix \ params id=Email
Group the primitives with their relevant identifiers in the correct order:
crm(live)configure#
group
g-mailsvc Public-IP Email
To change the order of a group member, use the
modgroup
command from the
configure
subcommand. Use the following commands to
move the primitive Email
before
Public-IP
. (This is just to demonstrate the feature):
crm(live)configure#
modgroup
g-mailsvc add Email before Public-IP
In case you want to remove a resource from a group (for example,
Email
), use this command:
crm(live)configure#
modgroup
g-mailsvc remove Email
For an overview, refer to Section 6.3.5.1, “Groups”.
Clones were initially conceived as a convenient way to start N instances of an IP resource and have them distributed throughout the cluster for load balancing. They have turned out to be useful for several purposes, including integrating with DLM, the fencing subsystem and OCFS2. You can clone any resource, provided the resource agent supports it.
Learn more about cloned resources in Section 6.3.5.2, “Clones”.
To create an anonymous clone resource, first create a primitive
resource and then refer to it with the clone
command. Do the following:
Log in as root
and start the crm
interactive shell:
root #
crm
configure
Configure the primitive, for example:
crm(live)configure#
primitive
Apache ocf:heartbeat:apache
Clone the primitive:
crm(live)configure#
clone
cl-apache Apache
Promotable clone resources (formerly known as multi-state resources) are a specialization of clones. This type allows the instances to be in one of two operating modes, be it active/passive, primary/secondary, or master/slave.
To create a promotable clone resource, first create a primitive resource and then the promotable clone resource. The promotable clone resource must support at least promote and demote operations.
Log in as root
and start the crm
interactive shell:
root #
crm
configure
Configure the primitive. Change the intervals if needed:
crm(live)configure#
primitive
my-rsc ocf:myCorp:myAppl \ op monitor interval=60 \ op monitor interval=61 role=Master
Create the promotable clone resource:
crm(live)configure#
ms
ms-rsc my-rsc
Apart from the possibility to configure your cluster resources, the
crm
tool also allows you to manage existing resources.
The following subsections gives you an overview.
When administering a cluster the command crm configure show
lists the current CIB objects like cluster configuration, global options,
primitives, and others:
root #
crm
configure show node 178326192: alice node 178326448: bob primitive admin_addr IPaddr2 \ params ip=192.168.2.1 \ op monitor interval=10 timeout=20 primitive stonith-sbd stonith:external/sbd \ params pcmk_delay_max=30 property cib-bootstrap-options: \ have-watchdog=true \ dc-version=1.1.15-17.1-e174ec8 \ cluster-infrastructure=corosync \ cluster-name=hacluster \ stonith-enabled=true \ placement-strategy=balanced \ standby-mode=true rsc_defaults rsc-options: \ resource-stickiness=1 \ migration-threshold=3 op_defaults op-options: \ timeout=600 \ record-pending=true
In case you have lots of resources, the output of show
is too verbose. To restrict the output, use the name of the resource.
For example, to list the properties of the primitive
admin_addr
only, append the resource name to
show
:
root #
crm
configure show admin_addr primitive admin_addr IPaddr2 \ params ip=192.168.2.1 \ op monitor interval=10 timeout=20
However, in some cases, you want to limit the output of specific resources
even more. This can be achieved with filters. Filters
limit the output to specific components. For example, to list the
nodes only, use type:node
:
root #
crm
configure show type:node node 178326192: alice node 178326448: bob
In case you are also interested in primitives, use the
or
operator:
root #
crm
configure show type:node or type:primitive node 178326192: alice node 178326448: bob primitive admin_addr IPaddr2 \ params ip=192.168.2.1 \ op monitor interval=10 timeout=20 primitive stonith-sbd stonith:external/sbd \ params pcmk_delay_max=30
Furthermore, to search for an object that starts with a certain string, use this notation:
root #
crm
configure show type:primitive and and 'admin*' primitive admin_addr IPaddr2 \ params ip=192.168.2.1 \ op monitor interval=10 timeout=20
To list all available types, enter crm configure show type:
and press the →| key. The Bash completion will give
you a list of all types.
To start a new cluster resource you need the respective identifier. Proceed as follows:
Log in as root
and start the crm
interactive shell:
root #
crm
Switch to the resource level:
crm(live)#
resource
Start the resource with start
and press the
→| key to show all known resources:
crm(live)resource#
start
ID
A resource will be automatically restarted if it fails, but each failure
raises the resource's failcount. If a
migration-threshold
has been set for that resource,
the node will no longer be allowed to run the resource when the
number of failures has reached the migration threshold.
Open a shell and log in as user root
.
Get a list of all your resources:
root #
crm
resource list ... Resource Group: dlm-clvm:1 dlm:1 (ocf:pacemaker:controld) Started clvm:1 (ocf:heartbeat:clvm) Started
To clean up the resource dlm
, for example:
root #
crm
resource cleanup dlm
Proceed as follows to remove a cluster resource:
Log in as root
and start the crm
interactive shell:
root #
crm
configure
Run the following command to get a list of your resources:
crm(live)#
resource
status
For example, the output can look like this (whereas myIP is the relevant identifier of your resource):
myIP (ocf:IPaddr:heartbeat) ...
Delete the resource with the relevant identifier (which implies a
commit
too):
crm(live)#
configure
delete YOUR_ID
Commit the changes:
crm(live)#
configure
commit
Although resources are configured to automatically fail over (or migrate) to other nodes of the cluster if a hardware or software failure occurs, you can also manually move a resource to another node using either Hawk2 or the command line.
Use the migrate
command for this task. For example,
to migrate the resource ipaddress1
to a cluster node
named bob
, use these
commands:
root #
crm
resourcecrm(live)resource#
migrate
ipaddress1 bob
Tags are a way to refer to multiple resources at once, without creating
any colocation or ordering relationship between them. This can be useful
for grouping conceptually related resources. For example, if you have
several resources related to a database, create a tag called
databases
and add all resources related to the
database to this tag:
root #
crm
configure tag databases: db1 db2 db3
This allows you to start them all with a single command:
root #
crm
resource start databases
Similarly, you can stop them all too:
root #
crm
resource stop databases
The “health” status of a cluster or node can be displayed with so called scripts. A script can perform different tasks—they are not targeted to health. However, for this subsection, we focus on how to get the health status.
To get all the details about the health
command, use
describe
:
root #
crm
script describe health
It shows a description and a list of all parameters and their default
values. To execute a script, use run
:
root #
crm
script run health
If you prefer to run only one step from the suite, the
describe
command lists all available steps in the
Steps category.
For example, the following command executes the first step of the
health
command. The output is stored in the
health.json
file for further investigation:
root #
crm
script run health statefile='health.json'
It is also possible to run the above commands with
crm cluster health
.
For additional information regarding scripts, see http://crmsh.github.io/scripts/.
cib.xml
#Edit sourceIn case your cluster configuration contains sensitive information, such as passwords, it should be stored in local files. That way, these parameters will never be logged or leaked in support reports.
Before using secret
, better run the
show
command first to get an overview of all your
resources:
root #
crm
configure show primitive mydb ocf:heartbeat:mysql \ params replication_user=admin ...
To set a password for the above mydb
resource, use the following commands:
root #
crm
resource secret mydb set passwd linux INFO: syncing /var/lib/heartbeat/lrm/secrets/mydb/passwd to [your node list]
You can get the saved password back with:
root #
crm
resource secret mydb show passwd linux
Note that the parameters need to be synchronized between nodes; the
crm resource secret
command will take care of that. We
highly recommend to only use this command to manage secret parameters.
Investigating the cluster history is a complex task. To simplify this
task, crmsh contains the history
command with its
subcommands. It is assumed SSH is configured correctly.
Each cluster moves states, migrates resources, or starts important
processes. All these actions can be retrieved by subcommands of
history
.
By default, all history
commands look at the events of
the last hour. To change this time frame, use the
limit
subcommand. The syntax is:
root #
crm
historycrm(live)history#
limit
FROM_TIME [TO_TIME]
Some valid examples include:
limit
4:00pm
, limit
16:00
Both commands mean the same, today at 4pm.
limit
2012/01/12 6pm
January 12th 2012 at 6pm
limit
"Sun 5 20:46"
In the current year of the current month at Sunday the 5th at 8:46pm
Find more examples and how to create time frames at http://labix.org/python-dateutil.
The info
subcommand shows all the parameters which are
covered by the crm report
:
crm(live)history#
info
Source: live Period: 2012-01-12 14:10:56 - end Nodes: alice Groups: Resources:
To limit crm report
to certain parameters view the
available options with the subcommand help
.
To narrow down the level of detail, use the subcommand
detail
with a level:
crm(live)history#
detail
1
The higher the number, the more detailed your report will be. Default is
0
(zero).
After you have set above parameters, use log
to show
the log messages.
To display the last transition, use the following command:
crm(live)history#
transition
-1 INFO: fetching new logs, please wait ...
This command fetches the logs and runs dotty
(from the
graphviz package) to show the
transition graph. The shell opens the log file which you can browse with
the ↓ and ↑ cursor keys.
If you do not want to open the transition graph, use the
nograph
option:
crm(live)history#
transition
-1 nograph
The crm man page.
Visit the upstream project documentation at http://crmsh.github.io/documentation.
See Article “Highly Available NFS Storage with DRBD and Pacemaker” for an exhaustive example.
All tasks that need to be managed by a cluster must be available as a resource. There are two major groups here to consider: resource agents and STONITH agents. For both categories, you can add your own agents, extending the abilities of the cluster to your own needs.
A cluster sometimes detects that one of the nodes is behaving strangely and needs to remove it. This is called fencing and is commonly done with a STONITH resource.
It is impossible to know how SSH might react to other system problems.
For this reason, external SSH/STONITH agents (like
stonith:external/ssh
) are not supported for
production environments. If you still want to use such agents for
testing, install the
libglue-devel package.
To get a list of all currently available STONITH devices (from the
software side), use the command crm ra list stonith
.
If you do not find your favorite agent, install the
-devel package.
For more information on STONITH devices and resource agents,
see Chapter 10, Fencing and STONITH.
As of yet there is no documentation about writing STONITH agents. If you want to write new STONITH agents, consult the examples available in the source of the cluster-glue package.
All OCF resource agents (RAs) are available in
/usr/lib/ocf/resource.d/
, see
Section 6.3.2, “Supported Resource Agent Classes” for more information.
Each resource agent must supported the following operations to control
it:
start
start or enable the resource
stop
stop or disable the resource
status
returns the status of the resource
monitor
similar to status
, but checks also for unexpected
states
validate
validate the resource's configuration
meta-data
returns information about the resource agent in XML
The general procedure of how to create an OCF RA is like the following:
Load the file
/usr/lib/ocf/resource.d/pacemaker/Dummy
as a
template.
Create a new subdirectory for each new resource agents to avoid naming
contradictions. For example, if you have a resource group
kitchen
with the resource
coffee_machine
, add this resource to the directory
/usr/lib/ocf/resource.d/kitchen/
. To access this
RA, execute the command crm
:
root #
crm
configure primitive coffee_1 ocf:coffee_machine:kitchen ...
Implement the different shell functions and save your file under a different name.
More details about writing OCF resource agents can be found at http://linux-ha.org/wiki/Resource_Agents. Find special information about several concepts at Chapter 1, Product Overview.
According to the OCF specification, there are strict definitions of the exit codes an action must return. The cluster always checks the return code against the expected result. If the result does not match the expected value, then the operation is considered to have failed and a recovery action is initiated. There are three types of failure recovery:
Recovery Type |
Description |
Action Taken by the Cluster |
---|---|---|
soft |
A transient error occurred. |
Restart the resource or move it to a new location. |
hard |
A non-transient error occurred. The error may be specific to the current node. |
Move the resource elsewhere and prevent it from being retried on the current node. |
fatal |
A non-transient error occurred that will be common to all cluster nodes. This means a bad configuration was specified. |
Stop the resource and prevent it from being started on any cluster node. |
Assuming an action is considered to have failed, the following table outlines the different OCF return codes. It also shows the type of recovery the cluster will initiate when the respective error code is received.
OCF Return Code |
OCF Alias |
Description |
Recovery Type |
---|---|---|---|
0 |
OCF_SUCCESS |
Success. The command completed successfully. This is the expected result for all start, stop, promote and demote commands. |
soft |
1 |
OCF_ERR_GENERIC |
Generic “there was a problem” error code. |
soft |
2 |
OCF_ERR_ARGS |
The resource’s configuration is not valid on this machine (for example, it refers to a location/tool not found on the node). |
hard |
3 |
OCF_ERR_UNIMPLEMENTED |
The requested action is not implemented. |
hard |
4 |
OCF_ERR_PERM |
The resource agent does not have sufficient privileges to complete the task. |
hard |
5 |
OCF_ERR_INSTALLED |
The tools required by the resource are not installed on this machine. |
hard |
6 |
OCF_ERR_CONFIGURED |
The resource’s configuration is invalid (for example, required parameters are missing). |
fatal |
7 |
OCF_NOT_RUNNING |
The resource is not running. The cluster will not attempt to stop a resource that returns this for any action.
This OCF return code may or may not require resource
recovery—it depends on what is the expected resource status.
If unexpected, then |
N/A |
8 |
OCF_RUNNING_MASTER |
The resource is running in Master mode. |
soft |
9 |
OCF_FAILED_MASTER |
The resource is in Master mode but has failed. The resource will be demoted, stopped and then started (and possibly promoted) again. |
soft |
other |
N/A |
Custom error code. |
soft |
Fencing is a very important concept in computer clusters for HA (High Availability). A cluster sometimes detects that one of the nodes is behaving strangely and needs to remove it. This is called fencing and is commonly done with a STONITH resource. Fencing may be defined as a method to bring an HA cluster to a known state.
Every resource in a cluster has a state attached. For example: “resource r1 is started on alice”. In an HA cluster, such a state implies that “resource r1 is stopped on all nodes except alice”, because the cluster must make sure that every resource may be started on only one node. Every node must report every change that happens to a resource. The cluster state is thus a collection of resource states and node states.
When the state of a node or resource cannot be established with certainty, fencing comes in. Even when the cluster is not aware of what is happening on a given node, fencing can ensure that the node does not run any important resources.
There are two classes of fencing: resource level and node level fencing. The latter is the primary subject of this chapter.
Resource level fencing ensures exclusive access to a given resource. Common examples of this are changing the zoning of the node from a SAN fiber channel switch (thus locking the node out of access to its disks) or methods like SCSI reserve. For examples, refer to Section 11.10, “Additional Mechanisms for Storage Protection”.
Node level fencing prevents a failed node from accessing shared resources entirely. This is usually done in a simple and abrupt way: reset or power off the node.
In a Pacemaker cluster, the implementation of node level fencing is STONITH
(Shoot The Other Node in the Head). The High Availability Extension
includes the stonith
command line tool, an extensible
interface for remotely powering down a node in the cluster. For an
overview of the available options, run stonith --help
or refer to the man page of stonith
for more
information.
To use node level fencing, you first need to have a fencing device. To get a list of STONITH devices which are supported by the High Availability Extension, run one of the following commands on any of the nodes:
root #
stonith -L
or
root #
crm ra list stonith
STONITH devices may be classified into the following categories:
Power Distribution Units are an essential element in managing power capacity and functionality for critical network, server and data center equipment. They can provide remote load monitoring of connected equipment and individual outlet power control for remote power recycling.
A stable power supply provides emergency power to connected equipment by supplying power from a separate source if a utility power failure occurs.
If you are running a cluster on a set of blades, then the power control device in the blade enclosure is the only candidate for fencing. Of course, this device must be capable of managing single blade computers.
Lights-out devices (IBM RSA, HP iLO, Dell DRAC) are becoming increasingly popular and may even become standard in off-the-shelf computers. However, they are inferior to UPS devices, because they share a power supply with their host (a cluster node). If a node stays without power, the device supposed to control it would be useless. In that case, the CRM would continue its attempts to fence the node indefinitely while all other resource operations would wait for the fencing/STONITH operation to complete.
Testing devices are used exclusively for testing purposes. They are usually more gentle on the hardware. Before the cluster goes into production, they must be replaced with real fencing devices.
The choice of the STONITH device depends mainly on your budget and the kind of hardware you use.
The STONITH implementation of SUSE® Linux Enterprise High Availability Extension consists of two components:
pacemaker-fenced
is a daemon which can be accessed by local processes or over
the network. It accepts the commands which correspond to fencing
operations: reset, power-off, and power-on. It can also check the
status of the fencing device.
The pacemaker-fenced
daemon runs on every node in the High Availability cluster. The
pacemaker-fenced
instance running on the DC node receives a fencing request
from the pacemaker-controld
. It
is up to this and other pacemaker-fenced
programs to carry
out the desired fencing operation.
For every supported fencing device there is a STONITH plug-in which
is capable of controlling said device. A STONITH plug-in is the
interface to the fencing device. The STONITH plug-ins contained in
the cluster-glue package reside in
/usr/lib64/stonith/plugins
on each node.
(If you installed the
fence-agents package, too,
the plug-ins contained there are installed in
/usr/sbin/fence_*
.) All STONITH plug-ins look
the same to pacemaker-fenced
,
but are quite different on the other side, reflecting the nature of the
fencing device.
Some plug-ins support more than one device. A typical example is
ipmilan
(or external/ipmi
)
which implements the IPMI protocol and can control any device which
supports this protocol.
To set up fencing, you need to configure one or more STONITH
resources—the pacemaker-fenced
daemon requires no configuration. All
configuration is stored in the CIB. A STONITH resource is a resource of
class stonith
(see
Section 6.3.2, “Supported Resource Agent Classes”). STONITH resources
are a representation of STONITH plug-ins in the CIB. Apart from the
fencing operations, the STONITH resources can be started, stopped and
monitored, like any other resource. Starting or stopping STONITH
resources means loading and unloading the STONITH device driver on a
node. Starting and stopping are thus only administrative operations and
do not translate to any operation on the fencing device itself. However,
monitoring does translate to logging it to the device (to verify that the
device will work in case it is needed). When a STONITH resource fails
over to another node it enables the current node to talk to the STONITH
device by loading the respective driver.
STONITH resources can be configured like any other resource. For details how to do so with your preferred cluster management tool:
The list of parameters (attributes) depends on the respective STONITH
type. To view a list of parameters for a specific device, use the
stonith
command:
stonith -t stonith-device-type -n
For example, to view the parameters for the ibmhmc
device type, enter the following:
stonith -t ibmhmc -n
To get a short help text for the device, use the -h
option:
stonith -t stonith-device-type -h
In the following, find some example configurations written in the syntax
of the crm
command line tool. To apply them, put the
sample in a text file (for example, sample.txt
) and
run:
root #
crm
< sample.txt
For more information about configuring resources with the
crm
command line tool, refer to
Chapter 8, Configuring and Managing Cluster Resources (Command Line).
An IBM RSA lights-out device might be configured like this:
configure primitive st-ibmrsa-1 stonith:external/ibmrsa-telnet \ params nodename=alice ip_address=192.168.0.101 \ username=USERNAME password=PASSW0RD primitive st-ibmrsa-2 stonith:external/ibmrsa-telnet \ params nodename=bob ip_address=192.168.0.102 \ username=USERNAME password=PASSW0RD location l-st-alice st-ibmrsa-1 -inf: alice location l-st-bob st-ibmrsa-2 -inf: bob commit
In this example, location constraints are used for the following
reason: There is always a certain probability that the STONITH
operation is going to fail. Therefore, a STONITH operation on the
node which is the executioner as well is not reliable. If the node is
reset, it cannot send the notification about the fencing operation
outcome. The only way to do that is to assume that the operation is
going to succeed and send the notification beforehand. But if the
operation fails, problems could arise. Therefore, by convention,
pacemaker-fenced
refuses to terminate its host.
The configuration of a UPS type fencing device is similar to the examples above. The details are not covered here. All UPS devices employ the same mechanics for fencing. How the device is accessed varies. Old UPS devices only had a serial port, usually connected at 1200baud using a special serial cable. Many new ones still have a serial port, but often they also use a USB or Ethernet interface. The kind of connection you can use depends on what the plug-in supports.
For example, compare the apcmaster
with the
apcsmart
device by using the stonith
-t
stonith-device-type -n command:
stonith -t apcmaster -h
returns the following information:
STONITH Device: apcmaster - APC MasterSwitch (via telnet) NOTE: The APC MasterSwitch accepts only one (telnet) connection/session a time. When one session is active, subsequent attempts to connect to the MasterSwitch will fail. For more information see http://www.apc.com/ List of valid parameter names for apcmaster STONITH device: ipaddr login password For Config info [-p] syntax, give each of the above parameters in order as the -p value. Arguments are separated by white space. Config file [-F] syntax is the same as -p, except # at the start of a line denotes a comment
With
stonith -t apcsmart -h
you get the following output:
STONITH Device: apcsmart - APC Smart UPS (via serial port - NOT USB!). Works with higher-end APC UPSes, like Back-UPS Pro, Smart-UPS, Matrix-UPS, etc. (Smart-UPS may have to be >= Smart-UPS 700?). See http://www.networkupstools.org/protocols/apcsmart.html for protocol compatibility details. For more information see http://www.apc.com/ List of valid parameter names for apcsmart STONITH device: ttydev hostlist
The first plug-in supports APC UPS with a network port and telnet protocol. The second plug-in uses the APC SMART protocol over the serial line, which is supported by many APC UPS product lines.
Kdump belongs to the Special Fencing Devices and is in fact the opposite of a fencing device. The plug-in checks if a Kernel dump is in progress on a node. If so, it returns true, and acts as if the node has been fenced.
The Kdump plug-in must be used in concert with another, real STONITH
device, for example, external/ipmi
. For the fencing
mechanism to work properly, you must specify that Kdump is checked before
a real STONITH device is triggered. Use crm configure
fencing_topology
to specify the order of the fencing devices as
shown in the following procedure.
Use the stonith:fence_kdump
resource agent (provided
by the package fence-agents)
to monitor all nodes with the Kdump function enabled. Find a
configuration example for the resource below:
configure
primitive st-kdump stonith:fence_kdump \
params nodename="alice "\ 1
pcmk_host_check="static-list" \
pcmk_reboot_action="off" \
pcmk_monitor_action="metadata" \
pcmk_reboot_retries="1" \
timeout="60"
commit
Name of the node to be monitored. If you need to monitor more than one node, configure more STONITH resources. To prevent a specific node from using a fencing device, add location constraints. |
The fencing action will be started after the timeout of the resource.
In /etc/sysconfig/kdump
on each node, configure
KDUMP_POSTSCRIPT
to send a notification to all nodes
when the Kdump process is finished. For example:
/usr/lib/fence_kdump_send -i INTERVAL -p PORT -c 1 alice bob charlie [...]
The node that does a Kdump will restart automatically after Kdump has finished.
Write a new initrd
to include the library fence_kdump_send
with network enabled. Use the -f
option to overwrite
the existing file, so the new file will be used for the next boot process:
root #
dracut -f -a kdump
Open a port in the firewall for the fence_kdump
resource.
The default port is 7410
.
To achieve that Kdump is checked before triggering a real fencing
mechanism (like external/ipmi
),
use a configuration similar to the following:
fencing_topology \ alice: kdump-node1 ipmi-node1 \ bob: kdump-node2 ipmi-node2
For more details on fencing_topology
:
crm configure help fencing_topology
Like any other resource, the STONITH class agents also support the monitoring operation for checking status.
Monitor STONITH resources regularly, yet sparingly. For most devices a monitoring interval of at least 1800 seconds (30 minutes) should suffice.
Fencing devices are an indispensable part of an HA cluster, but the less you need to use them, the better. Power management equipment is often affected by too much broadcast traffic. Some devices cannot handle more than ten or so connections per minute. Some get confused if two clients try to connect at the same time. Most cannot handle more than one session at a time.
Checking the status of fencing devices once every few hours should usually be enough. The probability that a fencing operation needs to be performed and the power switch fails is low.
For detailed information on how to configure monitor operations, refer to Section 8.3.9, “Configuring Resource Monitoring” for the command line approach.
In addition to plug-ins which handle real STONITH devices, there are special purpose STONITH plug-ins.
Some STONITH plug-ins mentioned below are for demonstration and testing purposes only. Do not use any of the following devices in real-life scenarios because this may lead to data corruption and unpredictable results:
external/ssh
ssh
fence_kdump
This plug-in checks if a Kernel dump is in progress on a node. If so,
it returns true
, and acts as if the node has been
fenced. The node cannot run any resources during the dump anyway. This
avoids fencing a node that is already down but doing a dump, which
takes some time. The plug-in must be used in concert with another,
real STONITH device.
For configuration details, see Example 10.3, “Configuration of a Kdump Device”.
external/sbd
This is a self-fencing device. It reacts to a so-called “poison pill” which can be inserted into a shared disk. On shared-storage connection loss, it stops the node from operating. Learn how to use this STONITH agent to implement storage-based fencing in Chapter 11, Procedure 11.7, “Configuring the Cluster to Use SBD”. See also http://www.linux-ha.org/wiki/SBD_Fencing for more details.
external/sbd
and DRBD
The external/sbd
fencing mechanism requires that
the SBD partition is readable directly from each node. Thus, a DRBD*
device must not be used for an SBD partition.
However, you can use the fencing mechanism for a DRBD cluster, provided the SBD partition is located on a shared disk that is not mirrored or replicated.
external/ssh
Another software-based “fencing” mechanism. The nodes
must be able to log in to each other as root
without passwords.
It takes a single parameter, hostlist
, specifying
the nodes that it will target. As it is not able to reset a truly
failed node, it must not be used for real-life clusters—for
testing and demonstration purposes only. Using it for shared storage
would result in data corruption.
meatware
meatware
requires help from the user to operate.
Whenever invoked, meatware
logs a CRIT severity
message which shows up on the node's console. The operator then
confirms that the node is down and issues a
meatclient(8)
command. This tells
meatware
to inform the cluster that the node should
be considered dead. See
/usr/share/doc/packages/cluster-glue/README.meatware
for more information.
suicide
This is a software-only device, which can reboot a node it is running
on, using the reboot
command. This requires action
by the node's operating system and can fail under certain
circumstances. Therefore avoid using this device whenever possible.
However, it is safe to use on one-node clusters.
This configuration is useful if you want a fencing mechanism without shared storage. In this diskless mode, SBD fences nodes by using the hardware watchdog without relying on any shared device. However, diskless SBD cannot handle a split brain scenario for a two-node cluster. Use this option only for clusters with more than two nodes.
suicide
is the only exception to the “I do not shoot my host” rule.
Check the following list of recommendations to avoid common mistakes:
Do not configure several power switches in parallel.
To test your STONITH devices and their configuration, pull the plug once from each node and verify that fencing the node does takes place.
Test your resources under load and verify the timeout values are appropriate. Setting timeout values too low can trigger (unnecessary) fencing operations. For details, refer to Section 6.3.9, “Timeout Values”.
Use appropriate fencing devices for your setup. For details, also refer to Section 10.5, “Special Fencing Devices”.
Configure one or more STONITH resources. By default, the global
cluster option stonith-enabled
is set to
true
. If no STONITH resources have been defined,
the cluster will refuse to start any resources.
Do not set the global cluster option
stonith-enabled
to false
for the following reasons:
Clusters without STONITH enabled are not supported.
DLM/OCFS2 will block forever waiting for a fencing operation that will never happen.
Do not set the global cluster option
startup-fencing
to false
.
By default, it is set to true
for the following
reason: If a node is in an unknown state during cluster start-up, the
node will be fenced once to clarify its status.
/usr/share/doc/packages/cluster-glue
In your installed system, this directory contains README files for many STONITH plug-ins and devices.
Information about STONITH on the home page of The High Availability Linux Project.
Pacemaker Explained: Explains the concepts used to configure Pacemaker. Contains comprehensive and very detailed information for reference.
Article explaining the concepts of split brain, quorum and fencing in HA clusters.
SBD (STONITH Block Device) provides a node fencing mechanism for Pacemaker-based clusters through the exchange of messages via shared block storage (SAN, iSCSI, FCoE, etc.). This isolates the fencing mechanism from changes in firmware version or dependencies on specific firmware controllers. SBD needs a watchdog on each node to ensure that misbehaving nodes are really stopped. Under certain conditions, it is also possible to use SBD without shared storage, by running it in diskless mode.
The ha-cluster-bootstrap scripts provide an automated way to set up a cluster with the option of using SBD as fencing mechanism. For details, see the Article “Installation and Setup Quick Start”. However, manually setting up SBD provides you with more options regarding the individual settings.
This chapter explains the concepts behind SBD. It guides you through configuring the components needed by SBD to protect your cluster from potential data corruption in case of a split brain scenario.
In addition to node level fencing, you can use additional mechanisms for storage protection, such as LVM2 exclusive activation or OCFS2 file locking support (resource level fencing). They protect your system against administrative or application faults.
SBD expands to Storage-Based Death or STONITH Block Device.
The highest priority of the High Availability cluster stack is to protect the integrity of data. This is achieved by preventing uncoordinated concurrent access to data storage. The cluster stack takes care of this using several control mechanisms.
However, network partitioning or software malfunction could potentially cause scenarios where several DCs are elected in a cluster. If this so-called split brain scenario were allowed to unfold, data corruption might occur.
Node fencing via STONITH is the primary mechanism to prevent this. Using SBD as a node fencing mechanism is one way of shutting down nodes without using an external power off device in case of a split brain scenario.
In an environment where all nodes have access to shared storage, a small partition of the device is formatted for use with SBD. The size of the partition depends on the block size of the used disk (for example, 1 MB for standard SCSI disks with 512 byte block size or 4 MB for DASD disks with 4 kB block size). The initialization process creates a message layout on the device with slots for up to 255 nodes.
After the respective SBD daemon is configured, it is brought online on each node before the rest of the cluster stack is started. It is terminated after all other cluster components have been shut down, thus ensuring that cluster resources are never activated without SBD supervision.
The daemon automatically allocates one of the message slots on the partition to itself, and constantly monitors it for messages addressed to itself. Upon receipt of a message, the daemon immediately complies with the request, such as initiating a power-off or reboot cycle for fencing.
Also, the daemon constantly monitors connectivity to the storage device, and terminates itself in case the partition becomes unreachable. This guarantees that it is not disconnected from fencing messages. If the cluster data resides on the same logical unit in a different partition, this is not an additional point of failure: The workload will terminate anyway if the storage connectivity has been lost.
Whenever SBD is used, a correctly working watchdog is crucial. Modern systems support a hardware watchdog that needs to be “tickled” or “fed” by a software component. The software component (in this case, the SBD daemon) “feeds” the watchdog by regularly writing a service pulse to the watchdog. If the daemon stops feeding the watchdog, the hardware will enforce a system restart. This protects against failures of the SBD process itself, such as dying, or becoming stuck on an I/O error.
If Pacemaker integration is activated, SBD will not self-fence if device majority is lost. For example, your cluster contains three nodes: A, B, and C. Because of a network split, A can only see itself while B and C can still communicate. In this case, there are two cluster partitions: one with quorum because of being the majority (B, C), and one without (A). If this happens while the majority of fencing devices are unreachable, node A would immediately commit suicide, but nodes B and C would continue to run.
The following steps are necessary to manually set up storage-based protection.
They must be executed as root
. Before you start, check Section 11.3, “Requirements”.
Depending on your scenario, either use SBD with one to three devices or in diskless mode. For an outline, see Section 11.4, “Number of SBD Devices”. The detailed setup is described in:
You can use up to three SBD devices for storage-based fencing. When using one to three devices, the shared storage must be accessible from all nodes.
The path to the shared storage device must be persistent and
consistent across all nodes in the cluster. Use stable device names
such as /dev/disk/by-id/dm-uuid-part1-mpath-abcedf12345
.
The shared storage can be connected via Fibre Channel (FC), Fibre Channel over Ethernet (FCoE), or even iSCSI.
The shared storage segment must not use host-based RAID, LVM2, or DRBD*. DRBD can be split, which is problematic for SBD, as there cannot be two states in SBD. Cluster multi-device (Cluster MD) cannot be used for SBD.
However, using storage-based RAID and multipathing is recommended for increased reliability.
An SBD device can be shared between different clusters, as long as no more than 255 nodes share the device.
For clusters with more than two nodes, you can also use SBD in diskless mode.
SBD supports the use of up to three devices:
The most simple implementation. It is appropriate for clusters where all of your data is on the same shared storage.
This configuration is primarily useful for environments that use host-based mirroring but where no third storage device is available. SBD will not terminate itself if it loses access to one mirror leg, allowing the cluster to continue. However, since SBD does not have enough knowledge to detect an asymmetric split of the storage, it will not fence the other side while only one mirror leg is available. Thus, it cannot automatically tolerate a second failure while one of the storage arrays is down.
The most reliable configuration. It is resilient against outages of one device—be it because of failures or maintenance. SBD will terminate itself only if more than one device is lost and if required, depending on the status of the cluster partition or node. If at least two devices are still accessible, fencing messages can be successfully transmitted.
This configuration is suitable for more complex scenarios where storage is not restricted to a single array. Host-based mirroring solutions can have one SBD per mirror leg (not mirrored itself), and an additional tie-breaker on iSCSI.
This configuration is useful if you want a fencing mechanism without shared storage. In this diskless mode, SBD fences nodes by using the hardware watchdog without relying on any shared device. However, diskless SBD cannot handle a split brain scenario for a two-node cluster. Use this option only for clusters with more than two nodes.
When using SBD as a fencing mechanism, it is vital to consider the timeouts of all components, because they depend on each other.
This timeout is set during initialization of the SBD device. It depends mostly on your storage latency. The majority of devices must be successfully read within this time. Otherwise, the node might self-fence.
If your SBD device(s) reside on a multipath setup or iSCSI, the timeout should be set to the time required to detect a path failure and switch to the next path.
This also means that in /etc/multipath.conf
the
value of max_polling_interval
must be less than
watchdog
timeout.
msgwait
TimeoutThis timeout is set during initialization of the SBD device. It defines the time after which a message written to a node's slot on the SBD device is considered delivered. The timeout should be long enough for the node to detect that it needs to self-fence.
However, if the msgwait
timeout is relatively long,
a fenced cluster node might rejoin before the fencing action returns.
This can be mitigated by setting the SBD_DELAY_START
parameter in the SBD configuration, as described in
Procedure 11.4
in
Step 4.
stonith-timeout
in the CIBThis timeout is set in the CIB as a global cluster property. It defines how long to wait for the STONITH action (reboot, on, off) to complete.
stonith-watchdog-timeout
in the CIB
This timeout is set in the CIB as a global cluster property. If not set
explicitly, it defaults to 0
, which is appropriate for
using SBD with one to three devices. For use of SBD in diskless mode, see Procedure 11.8, “Configuring Diskless SBD” for more details.
If you change the watchdog timeout, you need to adjust the other two timeouts as well. The following “formula” expresses the relationship between these three values:
Timeout (msgwait) >= (Timeout (watchdog) * 2) stonith-timeout = Timeout (msgwait) + 20%
For example, if you set the watchdog timeout to 120
,
set the msgwait
timeout to 240
and the
stonith-timeout
to 288
.
If you use the ha-cluster-bootstrap scripts to set up a cluster and to initialize the SBD device, the relationship between these timeouts is automatically considered.
SUSE Linux Enterprise High Availability Extension ships with several kernel modules that provide hardware-specific watchdog drivers. For a list of the most commonly used ones, see Commonly Used Watchdog Drivers.
For clusters in production environments we recommend to use a hardware-specific
watchdog driver. However, if no watchdog matches your hardware,
softdog
can be used as kernel
watchdog module.
The High Availability Extension uses the SBD daemon as the software component that “feeds” the watchdog.
Finding the right watchdog kernel module for a given system is not trivial. Automatic probing fails very often. As a result, lots of modules are already loaded before the right one gets a chance.
Table 11.1
lists the most commonly used watchdog drivers. If your hardware is not listed there,
the directory
/lib/modules/KERNEL_VERSION/kernel/drivers/watchdog
gives you a list of choices, too. Alternatively, ask your hardware or
system vendor for details on system specific watchdog configuration.
Hardware | Driver |
---|---|
HP | hpwdt |
Dell, Lenovo (Intel TCO) | iTCO_wdt |
Fujitsu | ipmi_watchdog |
VM on z/VM on IBM mainframe | vmwatchdog |
Xen VM (DomU) | xen_xdt |
Generic | softdog |
Some hardware vendors ship systems management software that uses the watchdog for system resets (for example, HP ASR daemon). If the watchdog is used by SBD, disable such software. No other software must access the watchdog timer.
To make sure the correct watchdog module is loaded, proceed as follows:
List the drivers that have been installed with your kernel version:
root #
rpm
-ql kernel-VERSION |grep
watchdog
List any watchdog modules that are currently loaded in the kernel:
root #
lsmod
|egrep
"(wd|dog)"
If you get a result, unload the wrong module:
root #
rmmod
WRONG_MODULE
Enable the watchdog module that matches your hardware:
root #
echo
WATCHDOG_MODULE > /etc/modules-load.d/watchdog.confroot #
systemctl
restart systemd-modules-load
Test whether the watchdog module is loaded correctly:
root #
lsmod
|grep
dog
For clusters in production environments we recommend to use a hardware-specific watchdog
driver. However, if no watchdog matches your hardware, softdog
can be used as kernel watchdog module.
The softdog driver assumes that at least one CPU is still running. If all CPUs are stuck, the code in the softdog driver that should reboot the system will never be executed. In contrast, hardware watchdogs keep working even if all CPUs are stuck.
Enable the softdog driver:
root #
echo
softdog > /etc/modules-load.d/watchdog.conf
Add the softdog
module in /etc/modules-load.d/watchdog.conf
and restart a service:
root #
echo
softdog > /etc/modules-load.d/watchdog.confroot #
systemctl
restart systemd-modules-load
Test whether the softdog watchdog module is loaded correctly:
root #
lsmod
|grep
softdog
The following steps are necessary for setup:
Before you start, make sure the block device or devices you want to use for SBD meet the requirements specified in Section 11.3.
When setting up the SBD devices, you need to take several timeout values into account. For details, see Section 11.5, “Calculation of Timeouts”.
The node will terminate itself if the SBD daemon running on it has not updated the watchdog timer fast enough. After having set the timeouts, test them in your specific environment.
To use SBD with shared storage, you must first create the messaging
layout on one to three block devices. The sbd create
command
will write a metadata header to the specified device or devices. It will also
initialize the messaging slots for up to 255 nodes. If executed without any
further options, the command will use the default timeout settings.
Make sure the device or devices you want to use for SBD do not hold any
important data. When you execute the sbd create
command, roughly the first megabyte of the specified block devices
will be overwritten without further requests or backup.
Decide which block device or block devices to use for SBD.
Initialize the SBD device with the following command:
root #
sbd
-d /dev/SBD create
(Replace /dev/SBD
with your actual path name, for example:
/dev/disk/by-id/scsi-ST2000DM001-0123456_Wabcdefg
.)
To use more than one device for SBD, specify the -d
option multiple times, for
example:
root #
sbd
-d /dev/SBD1 -d /dev/SBD2 -d /dev/SBD3 create
If your SBD device resides on a multipath group, use the -1
and -4
options to adjust the timeouts to use for SBD. For
details, see Section 11.5, “Calculation of Timeouts”.
All timeouts are given in seconds:
root #
sbd
-d /dev/SBD -4 1801 -1 902 create
Check what has been written to the device:
root #
sbd
-d /dev/SBD dump Header version : 2.1 UUID : 619127f4-0e06-434c-84a0-ea82036e144c Number of slots : 255 Sector size : 512 Timeout (watchdog) : 5 Timeout (allocate) : 2 Timeout (loop) : 1 Timeout (msgwait) : 10 ==Header on disk /dev/SBD is dumped
As you can see, the timeouts are also stored in the header, to ensure that all participating nodes agree on them.
After you have initialized the SBD devices, edit the SBD configuration file, then enable and start the respective services for the changes to take effect.
Open the file /etc/sysconfig/sbd
.
Search for the following parameter: SBD_DEVICE.
It specifies the devices to monitor and to use for exchanging SBD messages.
Edit this line by replacing SBD with your SBD device:
SBD_DEVICE="/dev/SBD"
If you need to specify multiple devices in the first line, separate them with semicolons (the order of the devices does not matter):
SBD_DEVICE="/dev/SBD1; /dev/SBD2; /dev/SBD3"
If the SBD device is not accessible, the daemon will fail to start and inhibit cluster start-up.
Search for the following parameter: SBD_DELAY_START.
Enables or disables a delay. Set SBD_DELAY_START
to yes
if msgwait
is relatively
long, but your cluster nodes boot very fast.
Setting this parameter to yes
delays the start of
SBD on boot. This is sometimes necessary with virtual machines.
After you have added your SBD devices to the SBD configuration file,
enable the SBD daemon. The SBD daemon is a critical piece
of the cluster stack. It needs to be running when the cluster stack is running.
Thus, the sbd
service is started as a dependency whenever
the pacemaker
service is started.
On each node, enable the SBD service:
root #
systemctl
enable sbd
It will be started together with the Corosync service whenever the Pacemaker service is started.
Restart the cluster stack on each node:
root #
crm
cluster restart
This automatically triggers the start of the SBD daemon.
As a next step, test the SBD devices as described in Procedure 11.6.
The following command will dump the node slots and their current messages from the SBD device:
root #
sbd
-d /dev/SBD list
Now you should see all cluster nodes that have ever been started with SBD listed here.
For example, if you have a two-node cluster, the message slot should show
clear
for both nodes:
0 alice clear 1 bob clear
Try sending a test message to one of the nodes:
root #
sbd
-d /dev/SBD message alice test
The node will acknowledge the receipt of the message in the system log files:
May 03 16:08:31 alice sbd[66139]: /dev/SBD: notice: servant: Received command test from bob on disk /dev/SBD
This confirms that SBD is indeed up and running on the node and that it is ready to receive messages.
As a final step, you need to adjust the cluster configuration as described in Procedure 11.7.
To configure the use of SBD in the cluster, you need to do the following in the cluster configuration:
Set the stonith-timeout parameter to a value that matches your setting.
Configure the SBD STONITH resource.
For the calculation of the stonith-timeout refer to Section 11.5, “Calculation of Timeouts”.
Start a shell and log in as root
or equivalent.
Run crm
configure
.
Enter the following:
crm(live)configure#
property
stonith-enabled="true" 1crm(live)configure#
property
stonith-watchdog-timeout=0 2crm(live)configure#
property
stonith-timeout="40s" 3
This is the default configuration, because clusters without STONITH are not supported.
But in case STONITH has been deactivated for testing purposes,
make sure this parameter is set to | |
If not explicitly set, this value defaults to | |
A |
For a two-node cluster, decide if you want predictable or random delays. For other cluster setups you do not need to set this parameter.
This parameter enables a static delay before executing STONITH actions. It ensures that the nodes do not fence each other if separate fencing resources and different delay values are being used. The targeted node will loose in a “fencing race”. The parameter can be used to “mark” a specific node to survive in case of a split brain scenario in a two-node cluster. To make this succeed, it is essential to create two primitive STONITH devices for each node. In the following configuration, alice will win and survive in case of a split brain scenario:
crm(live)configure#
primitive
st-sbd-alice stonith:external/sbd params \ pcmk_host_list=alice pcmk_delay_base=20crm(live)configure#
primitive
st-sbd-bob stonith:external/sbd params \ pcmk_host_list=bob pcmk_delay_base=0
This parameter prevents double fencing when using slow devices such as SBD. It adds a random delay for STONITH actions on the fencing device. It is especially important for two-node clusters where otherwise both nodes might try to fence each other in case of a split brain scenario.
crm(live)configure#
primitive
stonith_sbd stonith:external/sbd params pcmk_delay_max=30
Review your changes with show
.
Submit your changes with commit
and leave the crm live
configuration with exit
.
After the resource has started, your cluster is successfully configured for use of SBD. It will use this method in case a node needs to be fenced.
SBD can be operated in a diskless mode. In this mode, a watchdog device will be used to reset the node in the following cases: if it loses quorum, if any monitored daemon is lost and not recovered, or if Pacemaker decides that the node requires fencing. Diskless SBD is based on “self-fencing” of a node, depending on the status of the cluster, the quorum and some reasonable assumptions. No STONITH SBD resource primitive is needed in the CIB.
Do not use diskless SBD as a fencing mechanism for two-node clusters. Use it only in clusters with three or more nodes. SBD in diskless mode cannot handle split brain scenarios for two-node clusters.
Open the file /etc/sysconfig/sbd
and use
the following entries:
SBD_PACEMAKER=yes SBD_STARTMODE=always SBD_DELAY_START=no SBD_WATCHDOG_DEV=/dev/watchdog SBD_WATCHDOG_TIMEOUT=5
The SBD_DEVICE
entry is not needed as no shared
disk is used. When this parameter is missing, the sbd
service does not start any watcher process for SBD devices.
On each node, enable the SBD service:
root #
systemctl
enable sbd
It will be started together with the Corosync service whenever the Pacemaker service is started.
Restart the cluster stack on each node:
root #
crm
cluster restart
This automatically triggers the start of the SBD daemon.
Check if the parameter have-watchdog=true has been automatically set:
root #
crm
configure show | grep have-watchdog have-watchdog=true
Run crm configure
and set the following cluster
properties on the crm shell:
crm(live)configure#
property
stonith-enabled="true" 1crm(live)configure#
property
stonith-watchdog-timeout=10 2
This is the default configuration, because clusters without STONITH are not supported.
But in case STONITH has been deactivated for testing purposes,
make sure this parameter is set to | |
For diskless SBD, this parameter must not equal zero.
It defines after how long it is assumed that the fencing target has already
self-fenced. Therefore its value needs to be >= the value of
|
Review your changes with show
.
Submit your changes with commit
and leave the crm live
configuration with exit
.
To test whether SBD works as expected for node fencing purposes, use one or all of the following methods:
To trigger a fencing action for node NODENAME:
root #
crm
node fence NODENAME
Check if the node is fenced and if the other nodes consider the node as fenced after the stonith-watchdog-timeout.
Identify the process ID of the SBD inquisitor:
root #
systemctl
status sbd ● sbd.service - Shared-storage based fencing daemon Loaded: loaded (/usr/lib/systemd/system/sbd.service; enabled; vendor preset: disabled) Active: active (running) since Tue 2018-04-17 15:24:51 CEST; 6 days ago Docs: man:sbd(8) Process: 1844 ExecStart=/usr/sbin/sbd $SBD_OPTS -p /var/run/sbd.pid watch (code=exited, status=0/SUCCESS) Main PID: 1859 (sbd) Tasks: 4 (limit: 4915) CGroup: /system.slice/sbd.service ├─1859 sbd: inquisitor [...]
Simulate an SBD failure by terminating the SBD inquisitor process.
In our example, the process ID of the SBD inquisitor is
1859
):
root #
kill
-9 1859
The node proactively self-fences. The other nodes notice the loss of the node and consider it has self-fenced after the stonith-watchdog-timeout.
With a normal configuration, a failure of a resource stop operation will trigger fencing. To trigger fencing manually, you can produce a failure of a resource stop operation. Alternatively, you can temporarily change the configuration of a resource monitor operation and produce a monitor failure as described below:
Configure an on-fail=fence
property for a resource monitor
operation:
op monitor interval=10 on-fail=fence
Let the monitoring operation fail (for example, by terminating the respective daemon, if the resource relates to a service).
This failure triggers a fencing action.
Apart from node fencing via STONITH there are other methods to achieve
storage protection at a resource level. For example, SCSI-3 and SCSI-4 use
persistent reservations whereas sfex
provides a locking
mechanism. Both methods are explained in the following subsections.
The SCSI specifications 3 and 4 define persistent reservations.
These are SCSI protocol features and can be used for I/O fencing and failover.
This feature is implemented in the sg_persist
Linux
command.
Any backing disks for sg_persist
must be SCSI
disk compatible. sg_persist
only works for devices like
SCSI disks or iSCSI LUNs.
Do not use it for IDE, SATA, or any block devices
which do not support the SCSI protocol.
Before you proceed, check if your disk supports persistent reservations. Use the following command (replace DISK with your device name):
root #
sg_persist
-n --in --read-reservation -d /dev/DISK
The result shows whether your disk supports persistent reservations:
Supported disk:
PR generation=0x0, there is NO reservation held
Unsupported disk:
PR in (Read reservation): command not supported Illegal request, Invalid opcode
If you get an error message (like the one above), replace the old disk with an SCSI compatible disk. Otherwise proceed as follows:
To create the primitive resource sg_persist
,
run the following commands as root
:
root #
crm
configurecrm(live)configure#
primitive
sg sg_persist \ params devs="/dev/sdc" reservation_type=3 \ op monitor interval=60 timeout=60
Add the sg_persist
primitive to a master-slave
group:
crm(live)configure#
ms
ms-sg sg \ meta master-max=1 notify=true
Do some tests. When the resource is in master/slave status, you can
mount and write on /dev/sdc1
on the cluster node where
the master instance is running, while you cannot write on the cluster node
where the slave instance is running.
Add a file system primitive for Ext4:
crm(live)configure#
primitive
ext4 ocf:heartbeat:Filesystem \ params device="/dev/sdc1" directory="/mnt/ext4" fstype=ext4
Add the following order relationship plus a collocation between the
sg_persist
master and the file system resource:
crm(live)configure#
order
o-ms-sg-before-ext4 inf: ms-sg:promote ext4:startcrm(live)configure#
colocation
col-ext4-with-sg-persist inf: ext4 ms-sg:Master
Check all your changes with the show
command.
Commit your changes.
For more information, refer to the sg_persist
man
page.
sfex
#Edit source
This section introduces sfex
, an additional low-level
mechanism to lock access to shared storage exclusively to one node. Note
that sfex does not replace STONITH. As sfex requires shared
storage, it is recommended that the SBD node fencing mechanism described
above is used on another partition of the storage.
By design, sfex cannot be used with workloads that require concurrency (such as OCFS2). It serves as a layer of protection for classic failover style workloads. This is similar to an SCSI-2 reservation in effect, but more general.
In a shared storage environment, a small partition of the storage is set aside for storing one or more locks.
Before acquiring protected resources, the node must first acquire the protecting lock. The ordering is enforced by Pacemaker. The sfex component ensures that even if Pacemaker were subject to a split brain situation, the lock will never be granted more than once.
These locks must also be refreshed periodically, so that a node's death does not permanently block the lock and other nodes can proceed.
In the following, learn how to create a shared partition for use with
sfex and how to configure a resource for the sfex lock in the CIB. A
single sfex partition can hold any number of locks, and needs 1 KB
of storage space allocated per lock.
By default, sfex_init
creates one lock on the partition.
The shared partition for sfex should be on the same logical unit as the data you want to protect.
The shared sfex partition must not use host-based RAID, nor DRBD.
Using an LVM2 logical volume is possible.
Create a shared partition for use with sfex. Note the name of this
partition and use it as a substitute for
/dev/sfex
below.
Create the sfex metadata with the following command:
root #
sfex_init
-n 1 /dev/sfex
Verify that the metadata has been created correctly:
root #
sfex_stat
-i 1 /dev/sfex ; echo $?
This should return 2
, since the lock is not
currently held.
The sfex lock is represented via a resource in the CIB, configured as follows:
crm(live)configure#
primitive
sfex_1 ocf:heartbeat:sfex \ # params device="/dev/sfex" index="1" collision_timeout="1" \ lock_timeout="70" monitor_interval="10" \ # op monitor interval="10s" timeout="30s" on-fail="fence"
To protect resources via an sfex lock, create mandatory ordering and
placement constraints between the resources to protect the sfex resource. If
the resource to be protected has the ID
filesystem1
:
crm(live)configure#
order
order-sfex-1 inf: sfex_1 filesystem1crm(live)configure#
colocation
col-sfex-1 inf: filesystem1 sfex_1
If using group syntax, add the sfex resource as the first resource to the group:
crm(live)configure#
group
LAMP sfex_1 filesystem1 apache ipaddr
The cluster administration tools like crm shell (crmsh) or
Hawk2 can be used by root
or any user in the group
haclient
. By default, these
users have full read/write access. To limit access or assign more
fine-grained access rights, you can use Access control
lists (ACLs).
Access control lists consist of an ordered set of access rules. Each rule allows read or write access or denies access to a part of the cluster configuration. Rules are typically combined to produce a specific role, then users may be assigned to a role that matches their tasks.
This ACL documentation only applies if your CIB is validated with the CIB
syntax version pacemaker-2.0
or higher. For details on
how to check this and upgrade the CIB version, see
Note: Upgrading the CIB Syntax Version.
If you have upgraded from SUSE Linux Enterprise High Availability Extension 11 SPx and kept your former CIB version, refer to the Access Control List chapter in the Administration Guide for SUSE Linux Enterprise High Availability Extension 11 SP3 or earlier. It is available from http://www.suse.com/documentation/.
Before you start using ACLs on your cluster, make sure the following conditions are fulfilled:
Ensure you have the same users on all nodes in your cluster, either by using NIS, Active Directory, or by manually adding the same users to all nodes.
All users for whom you want to modify access rights with ACLs must
belong to the haclient
group.
All users need to run crmsh by its absolute path
/usr/sbin/crm
.
If non-privileged users want to run crmsh, their
PATH
variable needs to be extended with
/usr/sbin
.
ACLs are an optional feature. By default, use of ACLs is disabled.
If ACLs are not enabled, root
and all users belonging to the
haclient
group have full
read/write access to the cluster configuration.
Even if ACLs are enabled and configured, both root
and the
default CRM owner hacluster
always have full access to the cluster
configuration.
To use ACLs you need some knowledge about XPath. XPath is a language for selecting nodes in an XML document. Refer to http://en.wikipedia.org/wiki/XPath or look into the specification at http://www.w3.org/TR/xpath/.
Before you can start configuring ACLs, you need to enable use of ACLs. To do so, use the following command in the crmsh:
root #
crm
configure property enable-acl=true
Alternatively, use Hawk2 as described in Procedure 12.1, “Enabling Use of ACLs with Hawk2”.
Log in to Hawk2:
https://HAWKSERVER:7630/
In the left navigation bar, select
to display the global cluster options and their current values.
Below No
.
Set its value to Yes
and apply your changes.
Access control lists consist of an ordered set of access rules. Each rule allows read or write access or denies access to a part of the cluster configuration. Rules are typically combined to produce a specific role, then users may be assigned to a role that matches their tasks. An ACL role is a set of rules which describe access rights to CIB. A rule consists of the following:
an access right like read
, write
,
or deny
a specification where to apply the rule. This specification can be a type, an ID reference, or an XPath expression.
Usually, it is convenient to bundle ACLs into roles and assign a specific role to system users (ACL targets). There are two methods to create ACL rules:
Section 12.3.1, “Setting ACL Rules via XPath Expressions”. You need to know the structure of the underlying XML to create ACL rules.
Section 12.3.2, “Setting ACL Rules via Abbreviations”. Create a shorthand syntax and ACL rules to apply to the matched objects.
To manage ACL rules via XPath, you need to know the structure of the underlying XML. Retrieve the structure with the following command that shows your cluster configuration in XML (see Example 12.1):
root #
crm
configure show xml
<num_updates="59" dc-uuid="175704363" crm_feature_set="3.0.9" validate-with="pacemaker-2.0" epoch="96" admin_epoch="0" cib-last-written="Fri Aug 8 13:47:28 2014" have-quorum="1"> <configuration> <crm_config> <cluster_property_set id="cib-bootstrap-options"> <nvpair name="stonith-enabled" value="true" id="cib-bootstrap-options-stonith-enabled"/> [...] </cluster_property_set> </crm_config> <nodes> <node id="175704363" uname="alice"/> <node id="175704619" uname="bob"/> </nodes> <resources> [...] </resources> <constraints/> <rsc_defaults> [...] </rsc_defaults> <op_defaults> [...] </op_defaults> <configuration> </cib>
With the XPath language you can locate nodes in this XML document. For
example, to select the root node (cib
) use the XPath
expression /cib
. To locate the global cluster
configurations, use the XPath expression
/cib/configuration/crm_config
.
As an example, Table 12.1, “Operator Role—Access Types and XPath Expressions” shows the parameters (access type and XPath expression) to create an “operator” role. Users with this role can only execute the tasks mentioned in the second column—they cannot reconfigure any resources (for example, change parameters or operations), nor change the configuration of colocation or ordering constraints.
Type |
XPath/Explanation |
---|---|
Write |
//crm_config//nvpair[@name='maintenance-mode'] Turn cluster maintenance mode on or off. |
Write |
//op_defaults//nvpair[@name='record-pending'] Choose whether pending operations are recorded. |
Write |
//nodes/node//nvpair[@name='standby'] Set node in online or standby mode. |
Write |
//resources//nvpair[@name='target-role'] Start, stop, promote or demote any resource. |
Write |
//resources//nvpair[@name='maintenance'] Select if a resource should be put to maintenance mode or not. |
Write |
//constraints/rsc_location Migrate/move resources from one node to another. |
Read |
/cib View the status of the cluster. |
For users who do not want to deal with the XML structure there is an easier method.
For example, consider the following XPath:
//*[@id="rsc1"]
which locates all the XML nodes with the ID rsc1
.
The abbreviated syntax is written like this:
ref:"rsc1"
This also works for constraints. Here is the verbose XPath:
//constraints/rsc_location
The abbreviated syntax is written like this:
type:"rsc_location"
The abbreviated syntax can be used in crmsh and Hawk2. The CIB daemon knows how to apply the ACL rules to the matching objects.
The following procedures show how to configure read-only access to the
cluster configuration by defining a monitor
role and
assigning it to a user. Alternatively, you can use crmsh to do so,
as described in Procedure 12.4, “Adding a Monitor Role and Assigning a User with crmsh”.
Log in to Hawk2:
https://HAWKSERVER:7630/
In the left navigation bar, select
.Click
.
Enter a unique monitor
.
As access Read
.
As /cib
.
Click
.
This creates a new role with the name monitor
, sets
the read
rights and applies this to all elements in
the CIB by using the XPath expression/cib
.
If necessary, add more rules by clicking the plus icon and specifying the respective parameters.
Sort the individual rules by using the arrow up or down buttons.
To assign the role we created in Procedure 12.2 to a system user (target), proceed as follows:
Log in to Hawk2:
https://HAWKSERVER:7630/
In the left navigation bar, select
.
To create a system user (ACL Target), click tux
.
Make sure this user belongs to the haclient
group.
To assign a role to the target, select one or multiple
.
In our example, select the monitor
role you created
in Procedure 12.2.
Confirm your choice.
To configure access rights for resources or constraints, you can also use the abbreviated syntax as explained in Section 12.3.2, “Setting ACL Rules via Abbreviations”.
The following procedure shows how to configure a read-only access to the
cluster configuration by defining a monitor
role and
assigning it to a user.
Log in as root
.
Start the interactive mode of crmsh:
root #
crm
configurecrm(live)configure#
Define your ACL role(s):
Use the role
command to define a new role:
crm(live)configure#
role
monitor read xpath:"/cib"
The previous command creates a new role with the name
monitor
, sets the read
rights
and applies it to all elements in the CIB by using the XPath
expression /cib
. If necessary, you can add more
access rights and XPath arguments.
Add additional roles as needed.
Assign your roles to one or multiple ACL targets, which are the
corresponding system users. Make sure they belong to the
haclient
group.
crm(live)configure#
acl_target
tux monitor
Check your changes:
crm(live)configure#
show
Commit your changes:
crm(live)configure#
commit
To configure access rights for resources or constraints, you can also use the abbreviated syntax as explained in Section 12.3.2, “Setting ACL Rules via Abbreviations”.
For many systems, it is desirable to implement network connections that comply to more than the standard data security or availability requirements of a typical Ethernet device. In these cases, several Ethernet devices can be aggregated to a single bonding device.
The configuration of the bonding device is done by means of bonding module
options. The behavior is determined through the mode of the bonding
device. By default, this is mode=active-backup
,
which means that a different slave device will become active if the active
slave fails.
When using Corosync, the bonding device is not managed by the cluster software. Therefore, the bonding device must be configured on each cluster node that might possibly need to access the bonding device.
To configure a bonding device, you need to have multiple Ethernet devices that can be aggregated to a single bonding device. Proceed as follows:
Start YaST as root
and select › .
In the
, switch to the tab, which shows the available devices.Check if the Ethernet devices to be aggregate to a bonding device have an IP address assigned. If yes, change it:
Select the device to change and click
.In the
tab of the dialog that opens, select the option .Click
to return to the tab in the dialog.To add a new bonding device:
Click
and set the to . Proceed with .Select how to assign the IP address to the bonding device. Three methods are at your disposal:
No Link and IP Setup (Bonding Slaves)
Dynamic Address (with DHCP or Zeroconf)
Statically assigned IP Address
Use the method that is appropriate for your environment. If Corosync manages virtual IP addresses, select
and assign an IP address to the interface.Switch to the
tab.It shows any Ethernet devices that have been configured as bonding slaves in Step 3.b. To select the Ethernet devices that you want to include into the bond, below activate the check box in front of the respective devices.
Edit the
. The following modes are available:balance-rr
Provides load balancing and fault tolerance, at the cost of out-of-order packet transmission. This may cause delays, for example, for TCP reassembly.
active-backup
Provides fault tolerance.
balance-xor
Provides load balancing and fault tolerance.
broadcast
Provides fault tolerance.
802.3ad
Provides dynamic link aggregation if supported by the connected switch.
balance-tlb
Provides load balancing for outgoing traffic.
balance-alb
Provides load balancing for incoming and outgoing traffic, if the network devices used allow the modifying of the network device's hardware address while in use.
Make sure to add the parameter miimon=100
to
. Without this parameter, the
link is not checked regularly, so the bonding driver might continue
to lose packets on a faulty link.
Click /etc/sysconfig/network/ifcfg-bondDEVICENUMBER
.
Sometimes it is necessary to replace a bonding slave interface with
another one, for example, if the respective network device constantly
fails. The solution is to set up hotplugging bonding slaves. It is also
necessary to change the udev
rules to match the
device by bus ID instead of by MAC address. This enables you to replace
defective hardware (a network card in the same slot but with a different
MAC address), if the hardware allows for that.
If you prefer manual configuration instead, refer to the SUSE Linux Enterprise Server SUSE Linux Enterprise High Availability Extension Administration Guide, chapter Basic Networking, section Hotplugging of Bonding Slaves.
Start YaST as root
and select › .
In the
, switch to the tab, which shows the already configured devices. If bonding slaves are already configured, the column shows it.For each of the Ethernet devices that have been aggregated to a bonding device, execute the following steps:
Select the device to change and click
. The dialog opens.
Switch to the On
Hotplug
.
Switch to the
tab.For the
, click and select the option.Click
and to return to the tab in the dialog. If you click the Ethernet device entry now, the bottom pane shows the device's details, including the bus ID.Click
to confirm your changes and leave the network settings.
At boot time, the network setup does not wait for the hotplug slaves, but
for the bond to become ready, which needs at least one available slave.
When one of the slave interfaces is removed from the system (unbind from
NIC driver, rmmod
of the NIC driver or true PCI
hotplug removal), the Kernel removes it from the bond automatically. When
a new card is added to the system (replacement of the hardware in the
slot), udev
renames it by applying the bus-based
persistent name rule and calls ifup
for it. The
ifup
call automatically joins it into the bond.
All modes and many options are explained in detail in the
/usr/src/linux/Documentation/networking/bonding.txt
after you have installed the package
kernel-source
.
For High Availability setups, the following options described therein are
especially important: miimon
and
use_carrier
.
Load Balancing makes a cluster of servers appear as one large, fast server to outside clients. This apparent single server is called a virtual server. It consists of one or more load balancers dispatching incoming requests and several real servers running the actual services. With a load balancing setup of High Availability Extension, you can build highly scalable and highly available network services, such as Web, cache, mail, FTP, media and VoIP services.
High Availability Extension supports two technologies for load balancing: Linux Virtual Server (LVS) and HAProxy. The key difference is Linux Virtual Server operates at OSI layer 4 (Transport), configuring the network layer of kernel, while HAProxy operates at layer 7 (Application), running in user space. Thus Linux Virtual Server needs fewer resources and can handle higher loads, while HAProxy can inspect the traffic, do SSL termination and make dispatching decisions based on the content of the traffic.
On the other hand, Linux Virtual Server includes two different software: IPVS (IP Virtual Server) and KTCPVS (Kernel TCP Virtual Server). IPVS provides layer 4 load balancing whereas KTCPVS provides layer 7 load balancing.
This section gives you a conceptual overview of load balancing in combination with high availability, then briefly introduces you to Linux Virtual Server and HAProxy. Finally, it points you to further reading.
The real servers and the load balancers may be interconnected by either high-speed LAN or by geographically dispersed WAN. The load balancers dispatch requests to the different servers. They make parallel services of the cluster appear as one virtual service on a single IP address (the virtual IP address or VIP). Request dispatching can use IP load balancing technologies or application-level load balancing technologies. Scalability of the system is achieved by transparently adding or removing nodes in the cluster.
High availability is provided by detecting node or service failures and reconfiguring the whole virtual server system appropriately, as usual.
There are several load balancing strategies. Here are some Layer 4 strategies, suitable for Linux Virtual Server:
Round Robin. The simplest strategy is to direct each connection to a different address, taking turns. For example, a DNS server can have several entries for a given host name. With DNS round robin, the DNS server will return all of them in a rotating order. Thus different clients will see different addresses.
Selecting the “best” server. Although this has several drawbacks, balancing could be implemented with an “the first server who responds” or “the least loaded server” approach.
Balance number of connections per server. A load balancer between users and servers can divide the number of users across multiple servers.
Geo Location. It is possible to direct clients to a server nearby.
Here are some Layer 7 strategies, suitable for HAProxy:
URI. Inspect the HTTP content and dispatch to a server most suitable for this specific URI.
URL parameter, RDP cookie. Inspect the HTTP content for a session parameter, possibly in post parameters, or the RDP (remote desktop protocol) session cookie, and dispatch to the server serving this session.
Although there is some overlap, HAProxy can be used in scenarios
where LVS/ipvsadm
is not adequate and vice versa:
SSL termination. The front-end load balancers can handle the SSL layer. Thus the cloud nodes do not need to have access to the SSL keys, or could take advantage of SSL accelerators in the load balancers.
Application level. HAProxy operates at the application level, allowing the load balancing decisions to be influenced by the content stream. This allows for persistence based on cookies and other such filters.
On the other hand, LVS/ipvsadm
cannot be fully
replaced by HAProxy:
LVS supports “direct routing”, where the load balancer is only in the inbound stream, whereas the outbound traffic is routed to the clients directly. This allows for potentially much higher throughput in asymmetric environments.
LVS supports stateful connection table replication (via
conntrackd
). This allows for
load balancer failover that is transparent to the client and server.
The following sections give an overview of the main LVS components and concepts. Then we explain how to set up Linux Virtual Server on High Availability Extension.
The main component of LVS is the ip_vs (or IPVS) Kernel code. It is part of the default Kernel and implements transport-layer load balancing inside the Linux Kernel (layer-4 switching). The node that runs a Linux Kernel including the IPVS code is called director. The IPVS code running on the director is the essential feature of LVS.
When clients connect to the director, the incoming requests are load-balanced across all cluster nodes: The director forwards packets to the real servers, using a modified set of routing rules that make the LVS work. For example, connections do not originate or terminate on the director, it does not send acknowledgments. The director acts as a specialized router that forwards packets from end users to real servers (the hosts that run the applications that process the requests).
The ldirectord
daemon is a
user space daemon for managing Linux Virtual Server and monitoring the real servers
in an LVS cluster of load balanced virtual servers. A configuration
file (see below) specifies the virtual services and their associated real servers and tells
ldirectord
how to configure the
server as an LVS redirector. When the daemon is initialized, it creates
the virtual services for the cluster.
By periodically requesting a known URL and checking the responses, the
ldirectord
daemon monitors the
health of the real servers. If a real server fails, it will be removed
from the list of available servers at the load balancer. When the
service monitor detects that the dead server has recovered and is
working again, it will add the server back to the list of available
servers. In case that all real servers should be down, a fall-back
server can be specified to which to redirect a Web service. Typically
the fall-back server is localhost, presenting an emergency page about
the Web service being temporarily unavailable.
The ldirectord
uses the
ipvsadm
tool (package
ipvsadm) to manipulate the
virtual server table in the Linux Kernel.
There are three different methods of how the director can send packets from the client to the real servers:
Incoming requests arrive at the virtual IP. They are forwarded to the real servers by changing the destination IP address and port to that of the chosen real server. The real server sends the response to the load balancer which in turn changes the destination IP address and forwards the response back to the client. Thus, the end user receives the replies from the expected source. As all traffic goes through the load balancer, it usually becomes a bottleneck for the cluster.
IP tunneling enables packets addressed to an IP address to be redirected to another address, possibly on a different network. The LVS sends requests to real servers through an IP tunnel (redirecting to a different IP address) and the real servers reply directly to the client using their own routing tables. Cluster members can be in different subnets.
Packets from end users are forwarded directly to the real server. The IP packet is not modified, so the real servers must be configured to accept traffic for the virtual server's IP address. The response from the real server is sent directly to the client. The real servers and load balancers need to be in the same physical network segment.
Deciding which real server to use for a new connection requested by a
client is implemented using different algorithms. They are available as
modules and can be adapted to specific needs. For an overview of
available modules, refer to the ipvsadm(8)
man page.
Upon receiving a connect request from a client, the director assigns a
real server to the client based on a schedule. The
scheduler is the part of the IPVS Kernel code which decides which real
server will get the next new connection.
More detailed description about Linux Virtual Server scheduling algorithms can be
found at http://kb.linuxvirtualserver.org/wiki/IPVS.
Furthermore, search for --scheduler
in the
ipvsadm
man page.
Related load balancing strategies for HAProxy can be found at http://www.haproxy.org/download/1.6/doc/configuration.txt.
You can configure Kernel-based IP load balancing with the YaST IP
Load Balancing module. It is a front-end for
ldirectord
.
To access the IP Load Balancing dialog, start YaST as root
and select › . Alternatively, start the YaST
cluster module as root
on a command line with
yast2 iplb
.
The default installation does not include the configuration file
/etc/ha.d/ldirectord.cf
.
This file is created by the YaST module. The tabs available in the
YaST module correspond to the structure of the
/etc/ha.d/ldirectord.cf
configuration file,
defining global options and defining the options for the virtual
services.
For an example configuration and the resulting processes between load balancers and real servers, refer to Example 14.1, “Simple ldirectord Configuration”.
If a certain parameter is specified in both the virtual server section and in the global section, the value defined in the virtual server section overrides the value defined in the global section.
The following procedure describes how to configure the most important
global parameters. For more details about the individual parameters
(and the parameters not covered here), click ldirectord
man
page.
With ldirectord
will connect to
each of the real servers to check if they are still online.
With
, set the time in which the real server should have responded after the last check.
With ldirectord
will attempt to
request the real servers until the check is considered failed.
With
define a timeout in seconds for negotiate checks.In
, enter the host name or IP address of the Web server onto which to redirect a Web service in case all real servers are down.If you want the system to send alerts in case the connection status to any real server changes, enter a valid e-mail address in
.With
, define after how many seconds the e-mail alert should be repeated if any of the real servers remains inaccessible.In
specify the server states for which e-mail alerts should be sent. If you want to define more than one state, use a comma-separated list.
With ldirectord
should continuously
monitor the configuration file for modification. If set to
yes
, the configuration is automatically reloaded
upon changes.
With the 0
which means that no new
connections will be accepted. Already established connections will
persist until they time out.
If you want to use an alternative path for logging, specify a path for
the log files in ldirectord
writes its log
files to /var/log/ldirectord.log
.
You can configure one or more virtual services by defining a couple of
parameters for each. The following procedure describes how to configure
the most important parameters for a virtual service. For more details
about the individual parameters (and the parameters not covered here),
click ldirectord
man page.
In the YaST IP Load Balancing module, switch to the
tab.a new virtual server or an existing virtual server. A new dialog shows the available options.
In VIP:port
services into one virtual service.
To specify the gate
, ipip
or
masq
, see
Section 14.2.3, “Packet Forwarding”.
Click the
button and enter the required arguments for each real server.
As Negotiate
.
If you have set the Negotiate
, you also need to define the type of
service to monitor. Select it from the
drop-down box.
In
, enter the URI to the object that is requested on each real server during the check intervals.If you want to check if the response from the real servers contains a certain string (“I'm alive” message), define a regular expression that needs to be matched. Enter the regular expression into . If the response from a real server contains this expression, the real server is considered to be alive.
Depending on the type of Step 6, you also need to
specify further parameters for authentication. Switch to the
tab and enter the details like
, ,
, or . For more
information, refer to the YaST help text or to the
ldirectord
man page.
Switch to the
tab.
Select the ipvsadm(8)
man page.
Select the tcp
or udp
. If the
virtual service is specified as a firewall mark, the protocol must be
fwm
.
Define further parameters, if needed. Confirm your configuration with
/etc/ha.d/ldirectord.cf
.
The values shown in Figure 14.1, “YaST IP Load Balancing—Global Parameters” and
Figure 14.2, “YaST IP Load Balancing—Virtual Services”, would lead to the following
configuration, defined in /etc/ha.d/ldirectord.cf
:
autoreload = yes 1 checkinterval = 5 2 checktimeout = 3 3 quiescent = yes 4 virtual = 192.168.0.200:80 5 checktype = negotiate 6 fallback = 127.0.0.1:80 7 protocol = tcp 8 real = 192.168.0.110:80 gate 9 real = 192.168.0.120:80 gate 9 receive = "still alive" 10 request = "test.html" 11 scheduler = wlc 12 service = http 13
Defines that | |
Interval in which | |
Time in which the real server should have responded after the last check. | |
Defines not to remove failed real servers from the Kernel's LVS
table, but to set their weight to | |
Virtual IP address (VIP) of the LVS. The LVS is available at port
| |
Type of check that should be performed to test if the real servers are still alive. | |
Server onto which to redirect a Web service all real servers for this service are down. | |
Protocol to be used. | |
Two real servers defined, both available at port
| |
Regular expression that needs to be matched in the response string from the real server. | |
URI to the object that is requested on each real server during the check intervals. | |
Selected scheduler to be used for load balancing. | |
Type of service to monitor. |
This configuration would lead to the following process flow: The
ldirectord
will connect to each
real server once every 5 seconds
(2)
and request 192.168.0.110:80/test.html
or
192.168.0.120:80/test.html
as specified in
9
and
11.
If it does not receive the expected still alive
string
(10)
from a real server within 3 seconds
(3)
of the last check, it will remove the real server from the available
pool. However, because of the quiescent=yes
setting
(4),
the real server will not be removed from the LVS table. Instead its
weight will be set to 0
so that no new connections
to this real server will be accepted. Already established connections
will be persistent until they time out.
Apart from the configuration of
ldirectord
with YaST, you
need to make sure the following conditions are fulfilled to complete the
LVS setup:
The real servers are set up correctly to provide the needed services.
The load balancing server (or servers) must be able to route traffic to the real servers using IP forwarding. The network configuration of the real servers depends on which packet forwarding method you have chosen.
To prevent the load balancing server (or servers) from becoming a
single point of failure for the whole system, you need to set up one
or several backups of the load balancer. In the cluster configuration,
configure a primitive resource for
ldirectord
, so that
ldirectord
can fail over to
other servers in case of hardware failure.
As the backup of the load balancer also needs the
ldirectord
configuration file
to fulfill its task, make sure the
/etc/ha.d/ldirectord.cf
is available on all
servers that you want to use as backup for the load balancer. You can
synchronize the configuration file with Csync2 as described in
Section 4.5, “Transferring the Configuration to All Nodes”.
The following section gives an overview of the HAProxy and how to set up on High Availability. The load balancer distributes all requests to its back-end servers. It is configured as active/passive, meaning if one master fails, the slave becomes the master. In such a scenario, the user will not notice any interruption.
In this section, we will use the following setup:
A load balancer, with the IP address
192.168.1.99
.
A virtual, floating IP address
192.168.1.99
.
Our servers (usually for Web content)
www.example1.com
(IP:
192.168.1.200
) and
www.example2.com
(IP:
192.168.1.201
)
To configure HAProxy, use the following procedure:
Install the haproxy package.
Create the file /etc/haproxy/haproxy.cfg
with the
following contents:
global 1 maxconn 256 daemon defaults 2 log global mode http option httplog option dontlognull retries 3 option redispatch maxconn 2000 timeout connect 5000 3 timeout client 50s 4 timeout server 50000 5 frontend LB bind 192.168.1.99:80 6 reqadd X-Forwarded-Proto:\ http default_backend LB backend LB mode http stats enable stats hide-version stats uri /stats stats realm Haproxy\ Statistics stats auth haproxy:password 7 balance roundrobin 8 option httpclose option forwardfor cookie LB insert option httpchk GET /robots.txt HTTP/1.0 server web1-srv 192.168.1.200:80 cookie web1-srv check server web2-srv 192.168.1.201:80 cookie web2-srv check
Section which contains process-wide and OS-specific options.
| |
Section which sets default parameters for all other sections following its declaration. Some important lines:
| |
The maximum time to wait for a connection attempt to a server to succeed. | |
The maximum time of inactivity on the client side. | |
The maximum time of inactivity on the server side. | |
Section which combines front-end and back-end sections in one.
| |
Credentials for HAProxy Statistic report page. | |
Load balancing will work in a round-robin process. |
Test your configuration file:
root #
haproxy
-f /etc/haproxy/haproxy.cfg -c
Add the following line to Csync2's configuration file
/etc/csync2/csync2.cfg
to make sure the
HAProxy configuration file is included:
include /etc/haproxy/haproxy.cfg
Synchronize it:
root #
csync2
-f /etc/haproxy/haproxy.cfgroot #
csync2
-xv
The Csync2 configuration part assumes that the HA nodes were
configured using ha-cluster-bootstrap
. For details,
see the Installation and Setup Quick Start.
Make sure HAProxy is disabled on both load balancers
(alice
and
bob
) as it is started by
Pacemaker:
root #
systemctl
disable haproxy
Configure a new CIB:
root #
crm
configurecrm(live)#
cib
new haproxy-configcrm(haproxy-config)#
primitive
haproxy systemd:haproxy \ op start timeout=120 interval=0 \ op stop timeout=120 interval=0 \ op monitor timeout=100 interval=5s \ meta target-role=Startedcrm(haproxy-config)#
primitive
vip IPaddr2 \ params ip=192.168.1.99 nic=eth0 cidr_netmask=23 broadcast=192.168.1.255 \ op monitor interval=5s timeout=120 on-fail=restartcrm(haproxy-config)#
group
g-haproxy vip haproxy
Verify the new CIB and correct any errors:
crm(haproxy-config)#
verify
Commit the new CIB:
crm(haproxy-config)#
cib
use livecrm(live)#
cib
commit haproxy-config
Project home page at http://www.linuxvirtualserver.org/.
For more information about ldirectord
, refer to its
comprehensive man page.
LVS Knowledge Base: http://kb.linuxvirtualserver.org/wiki/Main_Page
Apart from local clusters and metro area clusters, SUSE® Linux Enterprise High Availability Extension
15 SP1 also supports geographically dispersed clusters (Geo
clusters, sometimes also called multi-site clusters). That means you can
have multiple, geographically dispersed sites with a local cluster each.
Failover between these clusters is coordinated by a higher level entity,
the so-called booth
. For details on how to
use and set up Geo clusters, refer to Article “Geo Clustering Quick Start” and
Book “Geo Clustering Guide”.
To perform maintenance tasks on the cluster nodes, you might need to stop the resources running on that node, to move them, or to shut down or reboot the node. It might also be necessary to temporarily take over the control of resources from the cluster, or even to stop the cluster service while resources remain running.
This chapter explains how to manually take down a cluster node without negative side-effects. It also gives an overview of different options the cluster stack provides for executing maintenance tasks.
When you shut down or reboot a cluster node (or stop the Pacemaker service on a node), the following processes will be triggered:
The resources that are running on the node will be stopped or moved off the node.
If stopping the resources should fail or time out, the STONITH mechanism will fence the node and shut it down.
If your aim is to move the services off the node in an orderly fashion before shutting down or rebooting the node, proceed as follows:
On the node you want to reboot or shut down, log in as root
or
equivalent.
Put the node into standby
mode:
root #
crm -w node standby
That way, services can migrate off the node without being limited by the shutdown timeout of Pacemaker.
Check the cluster status with:
root #
crm status
It shows the respective node in standby
mode:
[...] Node bob: standby [...]
Stop the Pacemaker service on that node:
root #
crm cluster stop
Reboot the node.
To check if the node joins the cluster again:
Log in to the node as root
or equivalent.
Check if the Pacemaker service has started:
root #
crm cluster status
If not, start it:
root #
crm cluster start
Check the cluster status with:
root #
crm status
It should show the node coming online again.
Pacemaker offers a variety of options for performing system maintenance:
The global cluster property maintenance-mode
allows
you to put all resources into maintenance state at once. The cluster will
cease monitoring them and thus become oblivious to their status.
This option allows you to put all resources running on a specific node into maintenance state at once. The cluster will cease monitoring them and thus become oblivious to their status.
A node that is in standby mode can no longer run resources. Any resources
running on the node will be moved away or stopped (in case no other node
is eligible to run the resource). Also, all monitoring operations will be
stopped on the node (except for those with
role="Stopped"
).
You can use this option if you need to stop a node in a cluster while continuing to provide the services running on another node.
When this mode is enabled for a resource, no monitoring operations will be triggered for the resource.
Use this option if you need to manually touch the service that is managed by this resource and do not want the cluster to run any monitoring operations for the resource during that time.
The is-managed
meta attribute allows you to temporarily
“release” a resource from being managed by the cluster
stack. This means you can manually touch the service that is managed by
this resource (for example, to adjust any components). However, the
cluster will continue to monitor the resource and to
report any failures.
If you want the cluster to also cease monitoring the resource, use the per-resource maintenance mode instead (see Putting a Resource into Maintenance Mode).
If you need to do testing or maintenance work, follow the general steps below.
Otherwise you risk unwanted side effects, like resources not starting in an orderly fashion, unsynchronized CIBs across the cluster nodes, or even data loss.
Before you start, choose which of the options outlined in Section 16.2 is appropriate for your situation.
Apply this option with Hawk2 or crmsh.
Execute your maintenance task or tests.
After you have finished, put the resource, node or cluster back to “normal” operation.
To put the cluster in maintenance mode on the crm shell, use the following command:
root #
crm
configure property maintenance-mode=true
To put the cluster back to normal mode after your maintenance work is done, use the following command:
root #
crm
configure property maintenance-mode=false
Start a Web browser and log in to the cluster as described in Section 7.2, “Logging In”.
In the left navigation bar, select
.In the
group, select the attribute from the empty drop-down box and click the plus icon to add it.
To set maintenance-mode=true
, activate the check box
next to maintenance-mode
and confirm your changes.
After you have finished the maintenance task for the whole cluster,
deactivate the check box next to the maintenance-mode
attribute.
From this point on, High Availability Extension will take over cluster management again.
To put a node in maintenance mode on the crm shell, use the following command:
root #
crm
node maintenance NODENAME
To put the node back to normal mode after your maintenance work is done, use the following command:
root #
crm
node ready NODENAME
Start a Web browser and log in to the cluster as described in Section 7.2, “Logging In”.
In the left navigation bar, select
.In one of the individual nodes' views, click the wrench icon next to the node and select
.After you have finished your maintenance task, click the wrench icon next to the node and select
.To put a node in standby mode on the crm shell, use the following command:
root #
crm node standby NODENAME
To bring the node back online after your maintenance work is done, use the following command:
root #
crm node online NODENAME
Start a Web browser and log in to the cluster as described in Section 7.2, “Logging In”.
In the left navigation bar, select
.In one of the individual nodes' views, click the wrench icon next to the node and select
.Finish the maintenance task for the node.
To deactivate the standby mode, click the wrench icon next to the node and select
.To put a resource into maintenance mode on the crm shell, use the following command:
root #
crm
resource maintenance RESOURCE_ID true
To put the resource back into normal mode after your maintenance work is done, use the following command:
root #
crm
resource maintenance RESOURCE_ID false
Start a Web browser and log in to the cluster as described in Section 7.2, “Logging In”.
In the left navigation bar, select
.Select the resource you want to put in maintenance mode or unmanaged mode, click the wrench icon next to the resource and select
.Open the
category.From the empty drop-down list, select the
attribute and click the plus icon to add it.
Activate the check box next to maintenance
to set the
maintenance attribute to yes
.
Confirm your changes.
After you have finished the maintenance task for that resource, deactivate
the check box next to the maintenance
attribute for
that resource.
From this point on, the resource will be managed by the High Availability Extension software again.
To put a resource into unmanaged mode on the crm shell, use the following command:
root #
crm
resource unmanage RESOURCE_ID
To put it into managed mode again after your maintenance work is done, use the following command:
root #
crm
resource manage RESOURCE_ID
Start a Web browser and log in to the cluster as described in Section 7.2, “Logging In”.
From the left navigation bar, select
and go to the list.In the
column, click the arrow down icon next to the resource you want to modify and select .The resource configuration screen opens.
Below
, select the entry from the empty drop-down box.
Set its value to No
and click .
After you have finished your maintenance task, set
Yes
(which is the
default value) and apply your changes.
From this point on, the resource will be managed by the High Availability Extension software again.
If the cluster or a node is in maintenance mode, you can stop or restart cluster resources at will—the High Availability Extension will not attempt to restart them. If you stop the Pacemaker service on a node, all daemons and processes (originally started as Pacemaker-managed cluster resources) will continue to run.
If you attempt to start Pacemaker services on a node while the cluster or node is in maintenance mode, Pacemaker will initiate a single one-shot monitor operation (a “probe”) for every resource to evaluate which resources are currently running on that node. However, it will take no further action other than determining the resources' status.
If you want to take down a node while either the cluster or the node is in
maintenance mode
, proceed as follows:
On the node you want to reboot or shut down, log in as root
or
equivalent.
If you have a DLM resource (or other resources depending on DLM), make sure to explicitly stop those resources before stopping the Pacemaker service:
crm(live)resource#
stop RESOURCE_ID
The reason is that stopping Pacemaker also stops the Corosync service, on whose membership and messaging services DLM depends. If Corosync stops, the DLM resource will assume a split brain scenario and trigger a fencing operation.
Stop the Pacemaker service on that node:
root #
crm cluster stop
Shut down or reboot the node.
The Distributed Lock Manager (DLM) in the kernel is the base component used by OCFS2, GFS2, Cluster MD, and Cluster LVM (lvmlockd) to provide active-active storage at each respective layer.
Oracle Cluster File System 2 (OCFS2) is a general-purpose journaling file system that has been fully integrated since the Linux 2.6 Kernel. OCFS2 allows you to store application binary files, data files, and databases on devices on shared storage. All nodes in a cluster have concurrent read and write access to the file system. A user space control daemon, managed via a clone resource, provides the integration with the HA stack, in particular with Corosync and the Distributed Lock Manager (DLM).
Global File System 2 or GFS2 is a shared disk file system for Linux computer clusters. GFS2 allows all nodes to have direct concurrent access to the same shared block storage. GFS2 has no disconnected operating-mode, and no client or server roles. All nodes in a GFS2 cluster function as peers. GFS2 supports up to 32 cluster nodes. Using GFS2 in a cluster requires hardware to allow access to the shared storage, and a lock manager to control access to the storage.
SUSE recommends OCFS2 over GFS2 for your cluster environments if performance is one of your major requirements. Our tests have revealed that OCFS2 performs better as compared to GFS2 in such settings.
The distributed replicated block device (DRBD*) allows you to create a mirror of two block devices that are located at two different sites across an IP network. When used with Corosync, DRBD supports distributed high-availability Linux clusters. This chapter shows you how to install and set up DRBD.
When managing shared storage on a cluster, every node must be informed about changes that are done to the storage subsystem. The Logical Volume Manager 2 (LVM2), which is widely used to manage local storage, has been extended to support transparent management of volume groups across the whole cluster. Volume groups shared among multiple hosts can be managed using the same commands as local storage.
The cluster multi-device (Cluster MD) is a software based RAID
storage solution for a cluster. Currently, Cluster MD provides the redundancy of
RAID1 mirroring to the cluster. With SUSE Linux Enterprise High Availability Extension 15 SP1, RAID10 is
included as a technology preview. If you want to try RAID10, replace mirror
with 10
in the related mdadm
command.
This chapter shows you how to create and use Cluster MD.
A clustered Samba server provides a High Availability solution in your heterogeneous networks. This chapter explains some background information and how to set up a clustered Samba server.
Relax-and-Recover (“Rear”, in this chapter abbreviated as Rear) is a disaster recovery framework for use by system administrators. It is a collection of Bash scripts that need to be adjusted to the specific production environment that is to be protected in case of disaster.
No disaster recovery solution will work out-of-the-box. Therefore it is essential to take preparations before any disaster happens.
The Distributed Lock Manager (DLM) in the kernel is the base component used by OCFS2, GFS2, Cluster MD, and Cluster LVM (lvmlockd) to provide active-active storage at each respective layer.
To avoid single points of failure, redundant communication paths are important for High Availability clusters. This is also true for DLM communication. If network bonding (Link Aggregation Control Protocol, LACP) cannot be used for any reason, we highly recommend to define a redundant communication channel (a second ring) in Corosync. For details, see Procedure 4.3, “Defining a Redundant Communication Channel”.
Depending on the configuration in /etc/corosync/corosync.conf
, DLM then decides
whether to use the TCP or SCTP protocol for its communication:
If none
(which
means redundant ring configuration is disabled), DLM automatically uses
TCP. However, without a redundant communication channel, DLM communication
will fail if the TCP link is down.
If passive
(which
is the typical setting), and a second communication ring in /etc/corosync/corosync.conf
is configured correctly, DLM automatically uses SCTP. In this case, DLM
messaging has the redundancy capability provided by SCTP.
DLM uses the cluster membership services from Pacemaker which run in user space. Therefore, DLM needs to be configured as a clone resource that is present on each node in the cluster.
As OCFS2, GFS2, Cluster MD, and Cluster LVM (lvmlockd) all use DLM, it is enough to configure one resource for DLM. As the DLM resource runs on all nodes in the cluster it is configured as a clone resource.
If you have a setup that includes both OCFS2 and Cluster LVM, configuring one DLM resource for both OCFS2 and Cluster LVM is enough.
The configuration consists of a base group that includes several primitives and a base clone. Both base group and base clone can be used in various scenarios afterward (for both OCFS2 and Cluster LVM, for example). You only need to extend the base group with the respective primitives as needed. As the base group has internal colocation and ordering, this simplifies the overall setup as you do not need to specify several individual groups, clones and their dependencies.
Follow the steps below on one node in the cluster:
Start a shell and log in as root
or equivalent.
Run crm
configure
.
Enter the following to create the primitive resource for DLM:
crm(live)configure#
primitive
dlm ocf:pacemaker:controld \ op monitor interval="60" timeout="60"
Create a base group for the DLM resource and further storage-related resources:
crm(live)configure#
group
g-storage dlm
Clone the g-storage
group so that it runs on all nodes:
crm(live)configure#
clone
cl-storage g-storage \ meta interleave=true target-role=Started
Review your changes with show
.
If everything is correct, submit your changes with
commit
and leave the crm live configuration with
exit
.
Clusters without STONITH are not supported. If you set the global cluster
option stonith-enabled
to false
for
testing or troubleshooting purposes, the DLM resource and all services
depending on it (such as Cluster LVM, GFS2, and OCFS2) will fail to start.
Oracle Cluster File System 2 (OCFS2) is a general-purpose journaling file system that has been fully integrated since the Linux 2.6 Kernel. OCFS2 allows you to store application binary files, data files, and databases on devices on shared storage. All nodes in a cluster have concurrent read and write access to the file system. A user space control daemon, managed via a clone resource, provides the integration with the HA stack, in particular with Corosync and the Distributed Lock Manager (DLM).
OCFS2 can be used for the following storage solutions for example:
General applications and workloads.
Xen image store in a cluster. Xen virtual machines and virtual servers can be stored on OCFS2 volumes that are mounted by cluster servers. This provides quick and easy portability of Xen virtual machines between servers.
LAMP (Linux, Apache, MySQL, and PHP | Perl | Python) stacks.
As a high-performance, symmetric and parallel cluster file system, OCFS2 supports the following functions:
An application's files are available to all nodes in the cluster. Users simply install it once on an OCFS2 volume in the cluster.
All nodes can concurrently read and write directly to storage via the standard file system interface, enabling easy management of applications that run across the cluster.
File access is coordinated through DLM. DLM control is good for most cases, but an application's design might limit scalability if it contends with the DLM to coordinate file access.
Storage backup functionality is available on all back-end storage. An image of the shared application files can be easily created, which can help provide effective disaster recovery.
OCFS2 also provides the following capabilities:
Metadata caching.
Metadata journaling.
Cross-node file data consistency.
Support for multiple-block sizes up to 4 KB, cluster sizes up to 1 MB, for a maximum volume size of 4 PB (Petabyte).
Support for up to 32 cluster nodes.
Asynchronous and direct I/O support for database files for improved database performance.
OCFS2 is only supported by SUSE when used with the pcmk (Pacemaker) stack, as provided by SUSE Linux Enterprise High Availability Extension. SUSE does not provide support for OCFS2 in combination with the o2cb stack.
The OCFS2 Kernel module (ocfs2
) is installed
automatically in the High Availability Extension on SUSE® Linux Enterprise Server 15 SP1. To use
OCFS2, make sure the following packages are installed on each node in
the cluster: ocfs2-tools and
the matching ocfs2-kmp-*
packages for your Kernel.
The ocfs2-tools package provides the following utilities for management of OFS2 volumes. For syntax information, see their man pages.
OCFS2 Utility |
Description |
---|---|
debugfs.ocfs2 |
Examines the state of the OCFS file system for debugging. |
fsck.ocfs2 |
Checks the file system for errors and optionally repairs errors. |
mkfs.ocfs2 |
Creates an OCFS2 file system on a device, usually a partition on a shared physical or logical disk. |
mounted.ocfs2 |
Detects and lists all OCFS2 volumes on a clustered system. Detects and lists all nodes on the system that have mounted an OCFS2 device or lists all OCFS2 devices. |
tunefs.ocfs2 |
Changes OCFS2 file system parameters, including the volume label, number of node slots, journal size for all node slots, and volume size. |
Before you can create OCFS2 volumes, you must configure the following resources as services in the cluster: DLM and a STONITH resource.
The following procedure uses the crm
shell to
configure the cluster resources. Alternatively, you can also use
Hawk2 to configure the resources as described in
Section 18.6, “Configuring OCFS2 Resources With Hawk2”.
You need to configure a fencing device. Without a STONITH
mechanism (like external/sbd
) in place the
configuration will fail.
Start a shell and log in as root
or equivalent.
Create an SBD partition as described in Procedure 11.3, “Initializing the SBD Devices”.
Run crm
configure
.
Configure external/sbd
as fencing device with
/dev/sdb2
being a dedicated partition on the shared
storage for heartbeating and fencing:
crm(live)configure#
primitive
sbd_stonith stonith:external/sbd \ params pcmk_delay_max=30 meta target-role="Started"
Review your changes with show
.
If everything is correct, submit your changes with
commit
and leave the crm live configuration with
exit
.
For details on configuring the resource group for DLM, see Procedure 17.1, “Configuring a Base Group for DLM”.
After you have configured a DLM cluster resource as described in Section 18.3, “Configuring OCFS2 Services and a STONITH Resource”, configure your system to use OCFS2 and create OCFs2 volumes.
We recommend that you generally store application files and data files on different OCFS2 volumes. If your application volumes and data volumes have different requirements for mounting, it is mandatory to store them on different volumes.
Before you begin, prepare the block devices you plan to use for your OCFS2 volumes. Leave the devices as free space.
Then create and format the OCFS2 volume with the
mkfs.ocfs2
as described in
Procedure 18.2, “Creating and Formatting an OCFS2 Volume”. The most important parameters for the
command are listed in Table 18.2, “Important OCFS2 Parameters”.
For more information and the command syntax, refer to the
mkfs.ocfs2
man page.
OCFS2 Parameter |
Description and Recommendation |
---|---|
Volume Label ( |
A descriptive name for the volume to make it uniquely identifiable
when it is mounted on different nodes. Use the
|
Cluster Size ( |
Cluster size is the smallest unit of space allocated to a file to
hold the data. For the available options and recommendations, refer
to the |
Number of Node Slots ( |
The maximum number of nodes that can concurrently mount a volume. For each of the nodes, OCFS2 creates separate system files, such as the journals. Nodes that access the volume can be a combination of little-endian architectures (such as AMD64/Intel 64) and big-endian architectures (such as S/390x).
Node-specific files are called local files. A node slot
number is appended to the local file. For example:
Set each volume's maximum number of node slots when you create it,
according to how many nodes that you expect to concurrently mount
the volume. Use the
In case the |
Block Size ( |
The smallest unit of space addressable by the file system. Specify
the block size when you create the volume. For the available options
and recommendations, refer to the |
Specific Features On/Off ( |
A comma separated list of feature flags can be provided, and
For an overview of all available flags, refer to the
|
Pre-Defined Features ( |
Allows you to choose from a set of pre-determined file system
features. For the available options, refer to the
|
If you do not specify any features when creating and formatting
the volume with mkfs.ocfs2
, the following features are
enabled by default: backup-super
,
sparse
, inline-data
,
unwritten
, metaecc
,
indexed-dirs
, and xattr
.
Execute the following steps only on one of the cluster nodes.
Open a terminal window and log in as root
.
Check if the cluster is online with the command crm
status
.
Create and format the volume using the mkfs.ocfs2
utility. For information about the syntax for this command, refer to
the mkfs.ocfs2
man page.
For example, to create a new OCFS2 file system on
/dev/sdb1
that supports up to 32 cluster nodes,
enter the following commands:
root #
mkfs.ocfs2 -N 32 /dev/sdb1
You can either mount an OCFS2 volume manually or with the cluster manager, as described in Procedure 18.4, “Mounting an OCFS2 Volume with the Cluster Resource Manager”.
Open a terminal window and log in as root
.
Check if the cluster is online with the command crm
status
.
Mount the volume from the command line, using the
mount
command.
If you mount the OCFS2 file system manually for testing purposes, make sure to unmount it again before starting to use it by means of cluster resources.
To mount an OCFS2 volume with the High Availability software, configure an
ocfs2 file system resource in the cluster. The following procedure uses
the crm
shell to configure the cluster resources.
Alternatively, you can also use Hawk2 to configure the resources as
described in Section 18.6, “Configuring OCFS2 Resources With Hawk2”.
Start a shell and log in as root
or equivalent.
Run crm
configure
.
Configure Pacemaker to mount the OCFS2 file system on every node in the cluster:
crm(live)configure#
primitive
ocfs2-1 ocf:heartbeat:Filesystem \ params device="/dev/sdb1" directory="/mnt/shared" \ fstype="ocfs2" options="acl" \ op monitor interval="20" timeout="40" \ op start timeout="60" op stop timeout="60" \ meta target-role="Started"
Add the ocfs2-1
primitive
to the g-storage
group you created in
Procedure 17.1, “Configuring a Base Group for DLM”.
crm(live)configure#
modgroup
g-storage add ocfs2-1
The add
subcommand appends the new group
member by default. Because of the base group's internal colocation and ordering, Pacemaker
will only start the ocfs2-1
resource on nodes that also have a dlm
resource
already running.
Review your changes with show
.
If everything is correct, submit your changes with
commit
and leave the crm live configuration with
exit
.
Instead of configuring the DLM and the file system resource for OCFS2 manually with the crm shell, you can also use the OCFS2 template in Hawk2's
.The OCFS2 template in the not include the configuration of a STONITH resource. If you use the wizard, you still need to create an SBD partition on the shared storage and configure a STONITH resource as described in Procedure 18.1, “Configuring a STONITH Resource”.
doesUsing the OCFS2 template in the Hawk2 Procedure 17.1, “Configuring a Base Group for DLM” and Procedure 18.4, “Mounting an OCFS2 Volume with the Cluster Resource Manager”.
also leads to a slightly different resource configuration than the manual configuration described inLog in to Hawk2:
https://HAWKSERVER:7630/
In the left navigation bar, select
.
Expand the OCFS2 File System
.
Follow the instructions on the screen. If you need information about an option, click it to display a short help text in Hawk2. After the last configuration step,
the values you have entered.The wizard displays the configuration snippet that will be applied to the CIB and any additional changes, if required.
Check the proposed changes. If everything is according to your wishes, apply the changes.
A message on the screen shows if the action has been successful.
To use quotas on an OCFS2 file system, create and mount the files
system with the appropriate quota features or mount options,
respectively: ursquota
(quota for individual users) or
grpquota
(quota for groups). These features can also
be enabled later on an unmounted file system using
tunefs.ocfs2
.
When a file system has the appropriate quota feature enabled, it tracks
in its metadata how much space and files each user (or group) uses. Since
OCFS2 treats quota information as file system-internal metadata, you
do not need to run the quotacheck
(8) program. All
functionality is built into fsck.ocfs2 and the file system driver itself.
To enable enforcement of limits imposed on each user or group, run
quotaon
(8) like you would do for any other file
system.
For performance reasons each cluster node performs quota accounting
locally and synchronizes this information with a common central storage
once per 10 seconds. This interval is tuneable with
tunefs.ocfs2
, options
usrquota-sync-interval
and
grpquota-sync-interval
. Therefore quota information may
not be exact at all times and as a consequence users or groups can
slightly exceed their quota limit when operating on several cluster nodes
in parallel.
For more information about OCFS2, see the following links:
The OCFS2 project home page.
The former OCFS2 project home page at Oracle.
The project's former documentation home page.
Global File System 2 or GFS2 is a shared disk file system for Linux computer clusters. GFS2 allows all nodes to have direct concurrent access to the same shared block storage. GFS2 has no disconnected operating-mode, and no client or server roles. All nodes in a GFS2 cluster function as peers. GFS2 supports up to 32 cluster nodes. Using GFS2 in a cluster requires hardware to allow access to the shared storage, and a lock manager to control access to the storage.
SUSE recommends OCFS2 over GFS2 for your cluster environments if performance is one of your major requirements. Our tests have revealed that OCFS2 performs better as compared to GFS2 in such settings.
To use GFS2, make sure gfs2-utils and a matching gfs2-kmp-* package for your Kernel is installed on each node of the cluster.
The gfs2-utils package provides the following utilities for management of GFS2 volumes. For syntax information, see their man pages.
GFS2 Utility |
Description |
---|---|
fsck.gfs2 |
Checks the file system for errors and optionally repairs errors. |
gfs2_jadd |
Adds additional journals to a GFS2 file system. |
gfs2_grow |
Grow a GFS2 file system. |
mkfs.gfs2 |
Create a GFS2 file system on a device, usually a shared device or partition. |
tunegfs2 |
Allows viewing and manipulating the GFS2 file system parameters such
as |
Before you can create GFS2 volumes, you must configure DLM and a STONITH resource.
You need to configure a fencing device. Without a STONITH
mechanism (like external/sbd
) in place the
configuration will fail.
Start a shell and log in as root
or equivalent.
Create an SBD partition as described in Procedure 11.3, “Initializing the SBD Devices”.
Run crm
configure
.
Configure external/sbd
as fencing device with
/dev/sdb2
being a dedicated partition on the shared
storage for heartbeating and fencing:
crm(live)configure#
primitive
sbd_stonith stonith:external/sbd \ params pcmk_delay_max=30 meta target-role="Started"
Review your changes with show
.
If everything is correct, submit your changes with
commit
and leave the crm live configuration with
exit
.
For details on configuring the resource group for DLM, see Procedure 17.1, “Configuring a Base Group for DLM”.
After you have configured DLM as cluster resources as described in Section 19.2, “Configuring GFS2 Services and a STONITH Resource”, configure your system to use GFS2 and create GFS2 volumes.
We recommend that you generally store application files and data files on different GFS2 volumes. If your application volumes and data volumes have different requirements for mounting, it is mandatory to store them on different volumes.
Before you begin, prepare the block devices you plan to use for your GFS2 volumes. Leave the devices as free space.
Then create and format the GFS2 volume with the
mkfs.gfs2
as described in
Procedure 19.2, “Creating and Formatting a GFS2 Volume”. The most important parameters for the
command are listed in Table 19.2, “Important GFS2 Parameters”. For
more information and the command syntax, refer to the
mkfs.gfs2
man page.
GFS2 Parameter |
Description and Recommendation |
---|---|
Lock Protocol Name ( |
The name of the locking protocol to use. Acceptable locking protocols are lock_dlm (for shared storage) or if you are using GFS2 as a local file system (1 node only), you can specify the lock_nolock protocol. If this option is not specified, lock_dlm protocol will be assumed. |
Lock Table Name ( |
The lock table field appropriate to the lock module you are using.
It is
clustername:fsname.
clustername must match that in the
cluster configuration file, |
Number of Journals ( |
The number of journals for gfs2_mkfs to create. You need at least one journal per machine that will mount the file system. If this option is not specified, one journal will be created. |
Execute the following steps only on one of the cluster nodes.
Open a terminal window and log in as root
.
Check if the cluster is online with the command crm
status
.
Create and format the volume using the mkfs.gfs2
utility. For information about the syntax for this command, refer to
the mkfs.gfs2
man page.
For example, to create a new GFS2 file system on
/dev/sdb1
that supports up to 32 cluster nodes,
use the following command:
root #
mkfs.gfs2 -t hacluster:mygfs2 -p lock_dlm -j 32 /dev/sdb1
The hacluster
name relates to
the entry cluster_name
in the file
/etc/corosync/corosync.conf
(this is the default).
You can either mount a GFS2 volume manually or with the cluster manager, as described in Procedure 19.4, “Mounting a GFS2 Volume with the Cluster Manager”.
Open a terminal window and log in as root
.
Check if the cluster is online with the command crm
status
.
Mount the volume from the command line, using the
mount
command.
If you mount the GFS2 file system manually for testing purposes, make sure to unmount it again before starting to use it by means of cluster resources.
To mount a GFS2 volume with the High Availability software, configure an OCF file
system resource in the cluster. The following procedure uses the
crm
shell to configure the cluster resources.
Alternatively, you can also use Hawk2 to configure the resources.
Start a shell and log in as root
or equivalent.
Run crm
configure
.
Configure Pacemaker to mount the GFS2 file system on every node in the cluster:
crm(live)configure#
primitive
gfs2-1 ocf:heartbeat:Filesystem \ params device="/dev/sdb1" directory="/mnt/shared" fstype="gfs2" \ op monitor interval="20" timeout="40" \ op start timeout="60" op stop timeout="60" \ meta target-role="Stopped"
Create a base group that consists of the dlm
primitive you created in Procedure 17.1, “Configuring a Base Group for DLM” and the
gfs2-1
primitive. Clone the group:
crm(live)configure#
group
g-storage dlm gfs2-1clone
cl-storage g-storage \ meta interleave="true"
Because of the base group's internal colocation and ordering, Pacemaker
will only start the gfs2-1
resource on nodes that also have a dlm
resource
already running.
Review your changes with show
.
If everything is correct, submit your changes with
commit
and leave the crm live configuration with
exit
.
The distributed replicated block device (DRBD*) allows you to create a mirror of two block devices that are located at two different sites across an IP network. When used with Corosync, DRBD supports distributed high-availability Linux clusters. This chapter shows you how to install and set up DRBD.
DRBD replicates data on the primary device to the secondary device in a way that ensures that both copies of the data remain identical. Think of it as a networked RAID 1. It mirrors data in real-time, so its replication occurs continuously. Applications do not need to know that in fact their data is stored on different disks.
DRBD is a Linux Kernel module and sits between the I/O scheduler at the
lower end and the file system at the upper end, see
Figure 20.1, “Position of DRBD within Linux”. To communicate with DRBD, users
use the high-level command drbdadm
. For maximum
flexibility DRBD comes with the low-level tool
drbdsetup
.
The data traffic between mirrors is not encrypted. For secure data exchange, you should deploy a Virtual Private Network (VPN) solution for the connection.
DRBD allows you to use any block device supported by Linux, usually:
partition or complete hard disk
software RAID
Logical Volume Manager (LVM)
Enterprise Volume Management System (EVMS)
By default, DRBD uses the TCP ports 7788
and higher
for communication between DRBD nodes. Make sure that your firewall does
not prevent communication on the used ports.
You must set up the DRBD devices before creating file systems on them.
Everything pertaining to user data should be done solely via the
/dev/drbdN
device and
not on the raw device, as DRBD uses the last part of the raw device for
metadata. Using the raw device will cause inconsistent data.
With udev integration, you will also get symbolic links in the form
/dev/drbd/by-res/RESOURCES
which are easier to use and provide safety against misremembering the
devices' minor number.
For example, if the raw device is 1024 MB in size, the DRBD device has only 1023 MB available for data, with about 70 KB hidden and reserved for the metadata. Any attempt to access the remaining kilobytes via raw disks fails because it is not available for user data.
Install the High Availability Extension on both SUSE Linux Enterprise Server machines in your networked cluster as described in Part I, “Installation, Setup and Upgrade”. Installing High Availability Extension also installs the DRBD program files.
If you do not need the complete cluster stack but only want to use DRBD, install the packages drbd, drbd-kmp-FLAVOR, drbd-utils, and yast2-drbd.
To simplify the work with drbdadm
, use the Bash
completion support.
If you want to enable it in your current shell session, insert the
following command:
root #
source
/etc/bash_completion.d/drbdadm.sh
To use it permanently for root
, create, or extend a file
/root/.bashrc
and insert the previous line.
The following procedure uses the server names alice and bob,
and the cluster resource name r0
. It sets up
alice as the primary node and /dev/sda1
for
storage. Make sure to modify the instructions to use your own nodes and
file names.
The following sections assumes you have two nodes, alice
and bob, and that they should use the TCP port 7788
.
Make sure this port is open in your firewall.
Prepare your system:
Make sure the block devices in your Linux nodes are ready and partitioned (if needed).
If your disk already contains a file system that you do not need anymore, destroy the file system structure with the following command:
root #
dd
if=/dev/zero of=YOUR_DEVICE count=16 bs=1M
If you have more file systems to destroy, repeat this step on all devices you want to include into your DRBD setup.
If the cluster is already using DRBD, put your cluster in maintenance mode:
root #
crm
configure property maintenance-mode=true
If you skip this step when your cluster uses already DRBD, a syntax error in the live configuration will lead to a service shutdown.
As an alternative, you can also use
drbdadm
-c FILE
to
test a configuration file.
Configure DRBD by choosing your method:
If you have configured Csync2 (which should be the default), the DRBD configuration files are already included in the list of files need to be synchronized. To synchronize them, use:
root #
csync2
-xv /etc/drbd.d/
If you do not have Csync2 (or do not want to use it), copy the DRBD configuration files manually to the other node:
root #
scp
/etc/drbd.conf bob:/etc/root #
scp
/etc/drbd.d/* bob:/etc/drbd.d/
Perform the initial synchronization (see Section 20.3.3, “Initializing and Formatting DRBD Resource”).
Reset the cluster's maintenance mode flag:
root #
crm
configure property maintenance-mode=false
The DRBD9 feature “auto promote” can use a clone and file system resource instead of a master/slave connection. When using this feature while a file system is being mounted, DRBD will change to primary mode automatically.
The auto promote feature has currently restricted support. With DRBD 9, SUSE supports the same use cases that were also supported with DRBD 8. Use cases beyond that, such as setups with more than two nodes, are not supported.
To set up DRBD manually, proceed as follows:
Beginning with DRBD version 8.3, the former configuration file is
split into separate files, located under the directory
/etc/drbd.d/
.
Open the file /etc/drbd.d/global_common.conf
. It
contains already some global, pre-defined values. Go to the
startup
section and insert these lines:
startup { # wfc-timeout degr-wfc-timeout outdated-wfc-timeout # wait-after-sb; wfc-timeout 100; degr-wfc-timeout 120; }
These options are used to reduce the timeouts when booting, see https://docs.linbit.com/docs/users-guide-9.0/#ch-configure for more details.
Create the file /etc/drbd.d/r0.res
. Change the
lines according to your situation and save it:
resource r0 { 1 device /dev/drbd0; 2 disk /dev/sda1; 3 meta-disk internal; 4 on alice { 5 address 192.168.1.10:7788; 6 node-id 0; 7 } on bob { 5 address 192.168.1.11:7788; 6 node-id 1; 7 } disk { resync-rate 10M; 8 } connection-mesh { 9 hosts alice bob; } }
DRBD resource name that allows some association to the service that needs them.
For example, | |
The device name for DRBD and its minor number.
In the example above, the minor number 0 is used for DRBD. The udev
integration scripts will give you a symbolic link
| |
The raw device that is replicated between nodes. Note, in this
example the devices are the same on both nodes.
If you need different devices, move the | |
The meta-disk parameter usually contains the value
| |
The | |
The IP address and port number of the respective node. Each
resource needs an individual port, usually starting with
| |
The node ID is required when configuring more than two nodes. It is a unique, non-negative integer to distinguish the different nodes. | |
The synchronization rate. Set it to one third of the lower of the disk- and network bandwidth. It only limits the resynchronization, not the replication. | |
Defines all nodes of a mesh.
The |
Check the syntax of your configuration file(s). If the following command returns an error, verify your files:
root #
drbdadm
dump all
Continue with Section 20.3.3, “Initializing and Formatting DRBD Resource”.
YaST can be used to start with an initial setup of DRBD. After you have created your DRBD setup, you can fine-tune the generated files manually.
However, when you have changed the configuration files, do not use the YaST DRBD module anymore. The DRBD module supports only a limited set of basic configuration. If you use it again, it is very likely that the module will not show your changes.
To set up DRBD with YaST, proceed as follows:
Start YaST and select the configuration module *.YaSTsave
.
Leave the booting flag in off
); do not change that as
Pacemaker manages this service.
If you have a firewall running, enable
.Go to the Figure 20.2, “Resource Configuration”).
entry. Press to create a new resource (seeThe following parameters need to be set:
The name of the DRBD resource (mandatory)
The host name of the relevant node
The IP address and port number (default
7788
) for the respective
node
The block device path that is used to access the replicated data.
If the device contains a minor number, the associated block device
is usually named /dev/drbdX
, where
X is the device minor number. If the
device does not contain a minor number, make sure to add
minor 0
after the device name.
The raw device that is replicated between both nodes. If you use LVM, insert your LVM device name.
The internal
or specifies an explicit device
extended by an index to hold the meta data needed by DRBD.
A real device may also be used for multiple drbd resources. For
example, if your /dev/sda6[0]
for the first resource, you may
use /dev/sda6[1]
for the second resource.
However, there must be at least 128 MB space for each resource
available on this disk. The fixed metadata size limits the maximum
data size that you can replicate.
All of these options are explained in the examples in the
/usr/share/doc/packages/drbd/drbd.conf
file and in
the man page of drbd.conf(5)
.
Click
.Click
to enter the second DRBD resource and finish with .Close the resource configuration with
and .If you use LVM with DRBD, it is necessary to change some options in the LVM configuration file (see the
entry). This change can be done by the YaST DRBD module automatically.
The disk name of localhost for the DRBD resource and the default filter
will be rejected in the LVM filter. Only /dev/drbd
can be scanned for an LVM device.
For example, if /dev/sda1
is used as a DRBD disk,
the device name will be inserted as the first entry in the LVM filter.
To change the filter manually, click the
check box.
Save your changes with
.Continue with Section 20.3.3, “Initializing and Formatting DRBD Resource”.
After you have prepared your system and configured DRBD, initialize your disk for the first time:
On both nodes (alice and bob), initialize the meta data storage:
root #
drbdadm
create-md r0root #
drbdadm
up r0
To shorten the initial resynchronization of your DRBD resource check the following:
If the DRBD devices on all nodes have the same data (for example,
by destroying the file system structure with the
dd
command as shown in
Section 20.3, “Setting Up DRBD Service”), then skip the initial
resynchronization with the following command (on both nodes):
root #
drbdadm
new-current-uuid --clear-bitmap r0/0
The state will be Secondary/Secondary UpToDate/UpToDate
Otherwise, proceed with the next step.
On the primary node alice, start the resynchronization process:
root #
drbdadm
primary --force r0
Check the status with:
root #
drbdadm
status r0 r0 role:Primary disk:UpToDate bob role:Secondary peer-disk:UpToDate
Create your file system on top of your DRBD device, for example:
root #
mkfs.ext3
/dev/drbd0
Mount the file system and use it:
root #
mount
/dev/drbd0 /mnt/
Between DRBD 8 (shipped with SUSE Linux Enterprise High Availability Extension 12 SP1) and DRBD 9 (shipped with SUSE Linux Enterprise High Availability Extension 12 SP2), the metadata format has changed. DRBD 9 does not automatically convert previous metadata files to the new format.
After migrating to 12 SP2 and before starting DRBD, convert the DRBD
metadata to the version 9 format manually. To do so, use
drbdadm
create-md
. No configuration
needs to be changed.
With DRBD 9, SUSE supports the same use cases that were also supported with DRBD 8. Use cases beyond that, such as setups with more than two nodes, are not supported.
DRBD 9 will fall back to be compatible with version 8. For three nodes and more, you need to re-create the metadata to use DRBD version 9 specific options.
If you have a stacked DRBD resource, refer also to Section 20.5, “Creating a Stacked DRBD Device” for more information.
To keep your data and allow to add new nodes without re-creating new resources, do the following:
Set one node in standby mode.
Update all the DRBD packages on all of your nodes, see Section 20.2, “Installing DRBD Services”.
Add the new node information to your resource configuration:
node-id on every on
section.
connection-mesh section contains all host names in the hosts parameter.
See the example configuration in Procedure 20.1, “Manually Configuring DRBD”.
Enlarge the space of your DRBD disks when using internal
as meta-disk
key. Use a device that supports enlarging
the space like LVM.
As an alternative, change to an external disk for metadata
and use meta-disk DEVICE;
.
Re-create the metadata based on the new configuration:
root #
drbdadm
create-md RESOURCE
Cancel the standby mode.
A stacked DRBD device contains two other devices of which at least one device is also a DRBD resource. In other words, DRBD adds an additional node on top of an already existing DRBD resource (see Figure 20.3, “Resource Stacking”). Such a replication setup can be used for backup and disaster recovery purposes.
Three-way replication uses asynchronous (DRBD protocol A) and synchronous replication (DRBD protocol C). The asynchronous part is used for the stacked resource whereas the synchronous part is used for the backup.
Your production environment uses the stacked device. For example,
if you have a DRBD device /dev/drbd0
and a stacked
device /dev/drbd10
on top, the file system will
be created on /dev/drbd10
, see Example 20.1, “Configuration of a Three-Node Stacked DRBD Resource” for more details.
# /etc/drbd.d/r0.res resource r0 { protocol C; device /dev/drbd0; disk /dev/sda1; meta-disk internal; on amsterdam-alice { address 192.168.1.1:7900; } on amsterdam-bob { address 192.168.1.2:7900; } } resource r0-U { protocol A; device /dev/drbd10; stacked-on-top-of r0 { address 192.168.2.1:7910; } on berlin-charlie { disk /dev/sda10; address 192.168.2.2:7910; # Public IP of the backup node meta-disk internal; } }
When a DRBD replication link becomes interrupted, Pacemaker tries to promote the DRBD resource to another node. To prevent Pacemaker from starting a service with outdated data, enable resource-level fencing in the DRBD configuration file as shown in Example 20.2, “Configuration of DRBD with Resource-Level Fencing Using the Cluster Information Base (CIB)”.
resource RESOURCE { net { fencing resource-only; # ... } handlers { fence-peer "/usr/lib/drbd/crm-fence-peer.9.sh"; after-resync-target "/usr/lib/drbd/crm-unfence-peer.9.sh"; # ... } ... }
If the DRBD replication link becomes disconnected, DRBD does the following:
DRBD calls the crm-fence-peer.9.sh
script.
The script contacts the cluster manager.
The script determines the Pacemaker resource associated with this DRBD resource.
The script ensures that the DRBD resource no longer gets promoted to any other node. It stays on the currently active one.
If the replication link becomes connected again and DRBD completes its synchronization process, then the constraint is removed. The cluster manager is now free to promote the resource.
If the install and configuration procedures worked as expected, you are ready to run a basic test of the DRBD functionality. This test also helps with understanding how the software works.
Test the DRBD service on alice.
Open a terminal console, then log in as
root
.
Create a mount point on alice, such as
/srv/r0
:
root #
mkdir
-p /srv/r0
Mount the drbd
device:
root #
mount
-o rw /dev/drbd0 /srv/r0
Create a file from the primary node:
root #
touch
/srv/r0/from_alice
Unmount the disk on alice:
root #
umount
/srv/r0
Downgrade the DRBD service on alice by typing the following command on alice:
root #
drbdadm
secondary r0
Test the DRBD service on bob.
Open a terminal console, then log in as root
on bob.
On bob, promote the DRBD service to primary:
root #
drbdadm
primary r0
On bob, check to see if bob is primary:
root #
drbdadm
status r0
On bob, create a mount point such as
/srv/r0
:
root #
mkdir
/srv/r0
On bob, mount the DRBD device:
root #
mount
-o rw /dev/drbd0 /srv/r0
Verify that the file you created on alice exists:
root #
ls
/srv/r0/from_alice
The /srv/r0/from_alice
file should be
listed.
If the service is working on both nodes, the DRBD setup is complete.
Set up alice as the primary again.
Dismount the disk on bob by typing the following command on bob:
root #
umount
/srv/r0
Downgrade the DRBD service on bob by typing the following command on bob:
root #
drbdadm
secondary r0
On alice, promote the DRBD service to primary:
root #
drbdadm
primary r0
On alice, check to see if alice is primary:
root #
drbdadm
status r0
To get the service to automatically start and fail over if the server has a problem, you can set up DRBD as a high availability service with Pacemaker/Corosync. For information about installing and configuring for SUSE Linux Enterprise 15 SP1 see Part II, “Configuration and Administration”.
DRBD comes with the utility drbdmon
which offers
realtime monitoring. It shows all the configured resources and their
problems.
drbdmon
#In case of problems, drbdadm
shows an error message:
drbdmon
#There are several ways to tune DRBD:
Use an external disk for your metadata. This might help, at the cost of maintenance ease.
Tune your network connection, by changing the receive and send buffer
settings via sysctl
.
Change the max-buffers
,
max-epoch-size
or both in the DRBD
configuration.
Increase the al-extents
value, depending on
your IO patterns.
If you have a hardware RAID controller with a BBU (Battery
Backup Unit), you might benefit from setting
no-disk-flushes
,
no-disk-barrier
and/or
no-md-flushes
.
Enable read-balancing depending on your workload. See https://www.linbit.com/en/read-balancing/ for more details.
The DRBD setup involves many components and problems may arise from different sources. The following sections cover several common scenarios and recommend various solutions.
If the initial DRBD setup does not work as expected, there is probably something wrong with your configuration.
To get information about the configuration:
Open a terminal console, then log in as root
.
Test the configuration file by running drbdadm
with
the -d
option. Enter the following command:
root #
drbdadm
-d adjust r0
In a dry run of the adjust
option,
drbdadm
compares the actual configuration of the
DRBD resource with your DRBD configuration file, but it does not
execute the calls. Review the output to make sure you know the source
and cause of any errors.
If there are errors in the /etc/drbd.d/*
and
drbd.conf
files, correct them before continuing.
If the partitions and settings are correct, run
drbdadm
again without the -d
option.
root #
drbdadm
adjust r0
This applies the configuration file to the DRBD resource.
For DRBD, host names are case-sensitive (Node0
would be a different host than node0
), and
compared to the host name as stored in the Kernel (see the
uname -n
output).
If you have several network devices and want to use a dedicated network
device, the host name will likely not resolve to the used IP address. In
this case, use the parameter disable-ip-verification
.
If your system cannot connect to the peer, this might be a problem with
your local firewall. By default, DRBD uses the TCP port
7788
to access the other node. Make sure that this
port is accessible on both nodes.
In cases when DRBD does not know which of the real devices holds the latest data, it changes to a split brain condition. In this case, the respective DRBD subsystems come up as secondary and do not connect to each other. In this case, the following message can be found in the logging data:
Split-Brain detected, dropping connection!
To resolve this situation, enter the following commands on the node which has data to be discarded:
root #
drbdadm
secondary r0
If the state is in WFconnection
, disconnect first:
root #
drbdadm
disconnect r0
On the node which has the latest data enter the following:
root #
drbdadm
connect --discard-my-data r0
That resolves the issue by overwriting one node's data with the peer's data, therefore getting a consistent view on both nodes.
The following open source resources are available for DRBD:
The project home page http://www.drbd.org.
See Article “Highly Available NFS Storage with DRBD and Pacemaker”.
http://clusterlabs.org/wiki/DRBD_HowTo_1.0 by the Linux Pacemaker Cluster Stack Project.
The following man pages for DRBD are available in the distribution:
drbd(8)
, drbdmeta(8)
,
drbdsetup(8)
, drbdadm(8)
,
drbd.conf(5)
.
Find a commented example configuration for DRBD at
/usr/share/doc/packages/drbd-utils/drbd.conf.example
.
Furthermore, for easier storage administration across your cluster, see the recent announcement about the DRBD-Manager at https://www.linbit.com/en/drbd-manager/.
When managing shared storage on a cluster, every node must be informed about changes that are done to the storage subsystem. The Logical Volume Manager 2 (LVM2), which is widely used to manage local storage, has been extended to support transparent management of volume groups across the whole cluster. Volume groups shared among multiple hosts can be managed using the same commands as local storage.
Cluster LVM is coordinated with different tools:
Coordinates access to shared resources among multiple hosts through cluster-wide locking.
LVM2 provides a virtual pool of disk space and enables flexible distribution of one logical volume over several disks.
The term Cluster LVM
indicates that LVM2 is being used
in a cluster environment. This needs some configuration adjustments
to protect the LVM2 metadata on shared storage. From SUSE Linux Enterprise 15 onward, the
cluster extension uses lvmlockd, which replaces the well-known
clvmd. For more information about lvmlockd, see the man page of the
lvmlockd
command (man 8
lvmlockd
).
Volume groups (VGs) and logical volumes (LVs) are basic concepts of LVM2. A volume group is a storage pool of multiple physical disks. A logical volume belongs to a volume group, and can be seen as an elastic volume on which you can create a file system. In a cluster environment, there is a concept of shared VGs, which consist of shared storage and can be used concurrently by multiple hosts.
Make sure the following requirements are fulfilled:
A shared storage device is available, provided by a Fibre Channel, FCoE, SCSI, iSCSI SAN, or DRBD*, for example.
Make sure the following packages have been installed: lvm2
and lvm2-lockd
.
From SUSE Linux Enterprise 15 onward, we use lvmlockd as the LVM2 cluster extension, rather than clvmd. Make sure the clvmd daemon is not running, otherwise lvmlockd will fail to start.
Perform the following basic steps on one node to configure a shared VG in the cluster:
Start a shell and log in as root
.
Check the current configuration of the cluster resources:
root #
crm configure show
If you have already configured a DLM resource (and a corresponding base group and base clone), continue with Procedure 21.2, “Creating an lvmlockd Resource”.
Otherwise, configure a DLM resource and a corresponding base group and base clone as described in Procedure 17.1, “Configuring a Base Group for DLM”.
Start a shell and log in as root
.
Run the following command to see the usage of this resource:
root #
crm configure ra info lvmlockd
Configure a lvmlockd
resource as follows:
root #
crm configure primitive lvmlockd ocf:heartbeat:lvmlockd \
op start timeout="90" \
op stop timeout="100" \
op monitor interval="30" timeout="90"
To ensure the lvmlockd
resource is started on every node, add the primitive resource
to the base group for storage you have created in Procedure 21.1, “Creating a DLM Resource”:
root #
crm configure modgroup g-storage add lvmlockd
Review your changes:
root #
crm configure show
Check if the resources are running well:
root #
crm status full
Start a shell and log in as root
.
Assuming you already have two shared disks, create a shared VG with them:
root #
vgcreate --shared vg1 /dev/sda /dev/sdb
Create an LV and do not activate it initially:
root #
lvcreate -an -L10G -n lv1 vg1
Start a shell and log in as root
.
Run the following command to see the usage of this resource:
root #
crm configure ra info LVM-activate
This resource manages the activation of a VG. In a shared VG, LV activation
has two different modes: exclusive and shared mode. The exclusive mode is
the default and should be used normally, when a local file system like ext4
uses the LV. The shared mode should only be used for cluster file systems
like OCFS2.
Configure a resource to manage the activation of your VG. Choose one of the following options according to your scenario:
Use exclusive activation mode for local file system usage:
root #
crm configure primitive vg1 ocf:heartbeat:LVM-activate \
params vgname=vg1 vg_access_mode=lvmlockd \
op start timeout=90s interval=0 \
op stop timeout=90s interval=0 \
op monitor interval=30s timeout=90s
Use shared activation mode for OCFS2 and add it to the cloned
g-storage
group:
root #
crm configure primitive vg1 ocf:heartbeat:LVM-activate \ params vgname=vg1 vg_access_mode=lvmlockd activation_mode=shared \ op start timeout=90s interval=0 \ op stop timeout=90s interval=0 \ op monitor interval=30s timeout=90sroot #
crm configure modgroup g-storage add vg1
Check if the resources are running well:
root #
crm status full
The following scenario uses two SAN boxes which export their iSCSI targets to several clients. The general idea is displayed in Figure 21.1, “Setup of a Shared Disk with Cluster LVM”.
The following procedures will destroy any data on your disks!
Configure only one SAN box first. Each SAN box needs to export its own iSCSI target. Proceed as follows:
Run YaST and click
› to start the iSCSI Server module.If you want to start the iSCSI target whenever your computer is booted, choose
, otherwise choose .If you have a firewall running, enable
.Switch to the
tab. If you need authentication, enable incoming or outgoing authentication or both. In this example, we select .Add a new iSCSI target:
Switch to the
tab.Click
.Enter a target name. The name needs to be formatted like this:
iqn.DATE.DOMAIN
For more information about the format, refer to Section 3.2.6.3.1. Type "iqn." (iSCSI Qualified Name) at http://www.ietf.org/rfc/rfc3720.txt.
If you want a more descriptive name, you can change it as long as your identifier is unique for your different targets.
Click
.Enter the device name in
and use a .Click
twice.Confirm the warning box with
.
Open the configuration file /etc/iscsi/iscsid.conf
and change the parameter node.startup
to
automatic
.
Now set up your iSCSI initiators as follows:
Run YaST and click
› .If you want to start the iSCSI initiator whenever your computer is booted, choose
, otherwise set .Change to the
tab and click the button.Add the IP address and the port of your iSCSI target (see Procedure 21.5, “Configuring iSCSI Targets (SAN)”). Normally, you can leave the port as it is and use the default value.
If you use authentication, insert the incoming and outgoing user name and password, otherwise activate
.Select
. The found connections are displayed in the list.Proceed with
.
Open a shell, log in as root
.
Test if the iSCSI initiator has been started successfully:
root #
iscsiadm
-m discovery -t st -p 192.168.3.100 192.168.3.100:3260,1 iqn.2010-03.de.jupiter:san1
Establish a session:
root #
iscsiadm
-m node -l -p 192.168.3.100 -T iqn.2010-03.de.jupiter:san1 Logging in to [iface: default, target: iqn.2010-03.de.jupiter:san1, portal: 192.168.3.100,3260] Login to [iface: default, target: iqn.2010-03.de.jupiter:san1, portal: 192.168.3.100,3260]: successful
See the device names with lsscsi
:
... [4:0:0:2] disk IET ... 0 /dev/sdd [5:0:0:1] disk IET ... 0 /dev/sde
Look for entries with IET
in their third column. In
this case, the devices are /dev/sdd
and
/dev/sde
.
Open a root
shell on one of the nodes you have run the iSCSI
initiator from
Procedure 21.6, “Configuring iSCSI Initiators”.
Create the shared volume group on disks /dev/sdd
and /dev/sde
:
root #
vgcreate --shared testvg /dev/sdd /dev/sde
Create logical volumes as needed:
root #
lvcreate
--name lv1 --size 500M testvg
Check the volume group with vgdisplay
:
--- Volume group --- VG Name testvg System ID Format lvm2 Metadata Areas 2 Metadata Sequence No 1 VG Access read/write VG Status resizable Clustered yes Shared no MAX LV 0 Cur LV 0 Open LV 0 Max PV 0 Cur PV 2 Act PV 2 VG Size 1016,00 MB PE Size 4,00 MB Total PE 254 Alloc PE / Size 0 / 0 Free PE / Size 254 / 1016,00 MB VG UUID UCyWw8-2jqV-enuT-KH4d-NXQI-JhH3-J24anD
After you have created the volumes and started your resources you should have new device
names under /dev/testvg
, for example /dev/testvg/lv1
.
This indicates the LV has been activated for use.